Development of a Dashboard for the Evolution of COVID-19

UNIVERSIDAD COMPLUTENSE DE MADRID FACULTAD DE INFORMÁTICA DEPARTAMENTO DE SISTEMAS INFORMÁTICOS Y COMPUTACIÓN T abajo Fin de G ado en Ingenie ía In o má ica Desa ollo de un Dashboa d pa a la E olución de la COVID-19 De elopmen o a Dashboa d o he E olu ion o COVID-19 Di igido po : Sonia Es é ez Ma ín Ad ián Ga cía Sánchez Ha old Luis Pascua Cajucom Dale F ancis Valencia Calicdan Cu so académico 2020-21 Con oca o ia Junio Acknowledgemen s Fi s o all, we wan o gi e hanks o ou p ojec di ec o , Sonia Es- é ez Ma in, who has always helped us when we ha e needed i and has gi en us guidelines o de elop he p ojec in he bes possible way. We also wan o men ion Vic o ia López, who has also helped us in planning he wo k and choosing adequa e me ics o he co ec p epa- a ion. We also wan o acknowledge all he e o s made on he pa o he pa ne g oup, composed o I án Jesús Sauga Peinado, Se gio Sánchez O iz, and Ál a o San José Gue a, who ha e p o ided us wi h he da ase s wi h which o wo k, a undamen al pa o he de elopmen o he p ojec and he achie emen o he objec i es. Finally, we wan o hank all amily and iends o all he mo al sup- po gi en h oughou his p ojec . 2 Abs ac The SARS-CoV-2 co ona i us is esponsible o he COVID- 19 disease ha has caused economic and heal h issues a ound he wo ld. To en o ce measu es ha cu bs he g ow h o his disease, his pandemic mus be moni o ed. The objec i e o his p ojec is o c ea e an online, open sou ce dashboa d ha moni o s he e olu ion o COVID-19 by coun y and by i s p incipal indica o s. To do so, his p ojec has de el- oped a se ies o Sc um sp in s, oge he wi h a ious echnologies, such as RS udio, Shiny, and Google Si es. The dashboa d c ea ed se es o analyze da a and o p esen i in a clea manne because o mas e y o di e en ools, me hod- ologies, and analy ical echniques ha ha e been used. This dashboa d di e s om o he exis ing dashboa ds such ha i specializes in p esen ing indica o s such as ansmission index, dange index, and cumula i e incidence in addi ion o aw da a like posi i e cases, dea hs, and eco e ies. As a esul , a dashboa d was c ea ed ha can in o m use s abou he p og ess o he pandemic in g aphic and map o ms, especially unde he a o emen ioned indica o s. Keywo ds COVID-19, Co ona i us, Dashboa d, RS udio, Shiny, Da a Analysis, Accu- mula ed Incidence, R0, Dange Index Resumen El SARS-CoV-2 co ona i us es esponsable de la COVID-19, la cual ha causado p oblemas económicos y sani a ios al ededo de odo el mundo. Pa a hace cumpli medidas que enen el c ecimien o de la en e medad, la pandemia debe se moni o izada. El obje i o de es e p oyec o es c ea un dashboa d online y de código abie o que moni o ice la e olución del COVID-19 po país y po sus p incipales indicado es. Pa a log a es o, el p oyec o ha sido elabo ado median e una se ie de i e aciones de Sc um, unidas a a ias ecnologías como RS udio, Shiny y Google Si es. El dashboa d c eado si e pa a analiza los da os y pa a p e- sen a los de una mane a cla a, debido al dominio de las di e en es he amien as, me odologías y écnicas analí icas que han sido u i- lizadas. Es e dashboa d di ie e de o os exis en es en que se especial- iza en p esen a algunos indicado es como son el índice de ans- misión, el índice de pelig osidad y la incidencia acumulada además de da os b u os como son los casos posi i os, las mue es y los e- cupe ados. Como esul ado, ha sido c eado un dashboa d que in o ma a los usua ios sob e el p og eso de la pandemia median e mapas y g á icos, especialmen e sob e los indicado es mencionados an e i- o men e. Palab as Cla e COVID-19, Co ona i us, Dashboa d, RS udio, Shiny, Análisis de da os, In- cidencia Acumulada, R0, Índice de pelig osidad Con en s 1 In oduc ion 8 1.1 Mo i a ion............................. 8 1.2 Objec i es............................. 9 1.3 Wo kPlan............................. 10 1.4 Documen S uc u e . . . . . . . . . . . . . . . . . . . . . . . 11 2 S a e o he A 13 2.1 Types o COVID-19 Dashboa ds . . . . . . . . . . . . . . . . . 13 2.1.1 Types o Dashboa ds by Geog aphic Scope . . . . . . . 14 2.1.2 Types o Dashboa ds by Type o Da a Used . . . . . . 17 2.2 Applica ions o Dashboa d C ea ion and Con igu a ion . . . . 21 2.3 Technical Requi emen s o C ea ing a Dashboa d . . . . . . . 23 3 Da a o In e es o he P ojec 25 3.1 Exis ing Da abases o In e es . . . . . . . . . . . . . . . . . . 25 3.2 G aphs o In e es . . . . . . . . . . . . . . . . . . . . . . . . . 26 3.3 Va iables and Indica o s o In e es . . . . . . . . . . . . . . . 31 3.3.1 Va iables.......................... 31 3.3.2 Indica o s ......................... 33 3.3.3 Smoo hed Va iables and Indica o s . . . . . . . . . . . 35 4 Da a P epa a ion 36 4.1 Da abases Used as Bases o Da abase C ea ion . . . . . . . . 36 4.2 Da ase C ea ion o P ojec Use . . . . . . . . . . . . . . . . 37 4.2.1 Fi s Da ase ....................... 37 4.2.2 Second Da ase . . . . . . . . . . . . . . . . . . . . . . 37 4.2.3 FinalDa ase ....................... 39 4.3 Da ase Quali y E alua ion . . . . . . . . . . . . . . . . . . . 39 4 CONTENTS 5 5 P ojec De elopmen 41 5.1 P ojec Planning ......................... 41 5.1.1 De elopmen Me hodology . . . . . . . . . . . . . . . . 41 5.1.2 GoogleSi es........................ 42 5.2 P oduc Backlog ......................... 43 5.3 Fi s Sp in ............................ 44 5.4 SecondSp in ........................... 49 5.5 Thi dSp in ............................ 55 5.6 Fou hSp in ........................... 59 5.7 RS udio Lib a ies Used . . . . . . . . . . . . . . . . . . . . . . 67 5.8 Sc um Sp in Schedule . . . . . . . . . . . . . . . . . . . . . . 68 6 Resul s 70 6.1 FinalDa ase ........................... 70 6.2 Dashboa d and Websi e . . . . . . . . . . . . . . . . . . . . . 72 6.3 Example Analysis: Compa ing Spain and he Philippines . . . 75 6.4 In eg a ing he Dashboa d o O he Websi es . . . . . . . . . 79 7 Indi idual Con ibu ions 80 7.1 Con ibu ions made by Ad ián Ga cía Sánchez . . . . . . . . . 80 7.2 Con ibu ions made by Dale F ancis Valencia Calicdan . . . . 82 7.3 Con ibu ions made by Ha old Luis Pascua Cajucom . . . . . 84 8 Conclusions and Fu u e Wo k 86 8.1 Conclusions ............................ 86 8.2 Fu u eWo k............................ 87 Lis o Figu es 2.1 ISCIII COVID-19 Dashboa d (Applica ion Used: Shiny) . . . 15 2.2 Co idly Dashboa d (Applica ion Used: Mapbox) . . . . . . . . 16 2.3 Global Map ound in he COVID-19 Dashboa d by JHU (Ap- plica ion Used: A cGIS Dashboa ds) . . . . . . . . . . . . . . 16 2.4 U.S. Map ound in he COVID-19 Dashboa d by JHU (Appli- ca ion Used: A cGIS Dashboa ds) . . . . . . . . . . . . . . . . 17 2.5 WHO COVID-19 Dashboa d (Applica ion Used: Sp inkl ) . . 18 2.6 Co id-19 T ialsT acke Dashboa d (Applica ion Used: Tableau) 19 2.7 Co id-T ials.o g Dashboa d (Applica ion Used: Cy el) . . . . 19 2.8 Se oT acke Dashboa d (Applica ions Used: Recha s and Lea le ) 20 2.9 The Vaccine T acke (Applica ions Used: Ga sby and Reac ) . 20 2.10 Typical a chi ec u e o a Dashboa d . . . . . . . . . . . . . . . 23 3.1 Rolling 7-day a e age o COVID-19 cases as o Oc obe 31, 2020 26 3.2 Table o con i med cases and dea hs o each coun y in he pas 14days............................ 27 3.3 Line g aph o he cumula i e con i med COVID-19 cases pe millionpeople........................... 28 3.4 Daily new con i med COVID-19 cases pe million people, Oc 31,2020 .............................. 29 3.5 BoxPlo .............................. 30 5.1 Sc um p ocess diag am . . . . . . . . . . . . . . . . . . . . . . 42 5.2 Ini ial designs (o de ed om op o bo om, le o igh ) . . . 45 5.3 Final p o o ype design. Map View 1 . . . . . . . . . . . . . . 46 5.4 Final p o o ype design. Map View 2 . . . . . . . . . . . . . . 46 5.5 Final p o o ype design. Map View 3 . . . . . . . . . . . . . . 47 5.6 Final p o o ype design. G aph View . . . . . . . . . . . . . . 47 6 LIST OF FIGURES 7 5.7 Sp in 1 esul (g aph iew) . . . . . . . . . . . . . . . . . . . 48 5.8 Sp in 1 esul (map iew) . . . . . . . . . . . . . . . . . . . . 48 5.9 Re ised p o o ype design (g aph iew) . . . . . . . . . . . . . 51 5.10 Re ised p o o ype design (map iew) . . . . . . . . . . . . . . 51 5.11 Sp in 2 esul (line g aph) . . . . . . . . . . . . . . . . . . . . 52 5.12 Sp in 2 esul (box plo ) . . . . . . . . . . . . . . . . . . . . . 53 5.13 Sp in 2 esul (map iew) . . . . . . . . . . . . . . . . . . . . 54 5.14 Re ised p o o ype design (g aph iew) . . . . . . . . . . . . . 56 5.15 Re ised p o o ype design (map iew) . . . . . . . . . . . . . . 56 5.16 Sp in 3 esul (g aph iew) . . . . . . . . . . . . . . . . . . . 58 5.17 Sp in 3 esul (map iew) . . . . . . . . . . . . . . . . . . . . 58 5.18 Dashboa d design (g aph iew) . . . . . . . . . . . . . . . . . 61 5.19 Dashboa d design (map iew) . . . . . . . . . . . . . . . . . . 62 5.20 Websi e design (home iew) . . . . . . . . . . . . . . . . . . . 63 5.21 Websi e implemen a ion (home iew) . . . . . . . . . . . . . . 65 5.22 Websi e implemen a ion(abou us iew) . . . . . . . . . . . . . 65 5.23 Websi e implemen a ion (abou his p ojec iew) . . . . . . . 66 5.24 Dashboa d a chi ec u e . . . . . . . . . . . . . . . . . . . . . . 68 5.25 Realized sp in da es . . . . . . . . . . . . . . . . . . . . . . . 69 6.1 Inc o Spain om July 2020 o May 2021, as shown in Dashboa d 72 6.2 Inc o Spain om July 2020 o May 2021 . . . . . . . . . . . . 72 6.3 Websi e wi h in eg a ed Dashboa d . . . . . . . . . . . . . . . 73 6.4 Func ions a ailable in he G aph View . . . . . . . . . . . . . 74 6.5 G aphs compa ing Spain and he Philippines unde R0 . . . . 76 6.6 Map compa ing Spain and he Philippines unde R0 on June 8,2020............................... 76 6.7 G aphs compa ing Spain and he Philippines unde DI . . . . 77 6.8 Map compa ing Spain and he Philippines unde DI on June 8,2020............................... 77 6.9 G aphs compa ing Spain and he Philippines unde Inc . . . . 78 6.10 Map compa ing Spain and he Philippines unde Inc on June 8,2020............................... 79 Chap e 1 In oduc ion COVID-19 [1] is an in ec ious disease caused by he co ona i us called Se- e e Acu e Respi a o y Synd ome Co ona i us 2 (SARS-CoV-2). I is held esponsible o he pandemic ha was happening a ound he wo ld decla ed by he Wo ld Heal h O ganiza ion (WHO) on Ma ch 11, 2020. An a icle published by WHO [2] sugges ed ha he s a o he ou b eak o his i us o igina ed om Wuhan, China a a local ma ke a ound la e Decembe 2019. E iden ly, he COVID-19 pandemic b ough de as a ing e ec s o e e y- one a ound he globe. Public heal h acili ies we e b ough o hei limi s [3], [4], economies wo ldwide ook a massi e hi [5], and, mos impo an ly, many li es ha e been los . WHO had s a ed ha , as o No embe 1, 2020, nea ly 46 million posi i e cases and 1.2 million dea hs ha e been epo ed om a ound he wo ld [6]. As such, he global phenomenon has challenged e e y s a e and coun y o o mula e hei own coun e measu es and impede he a e a which ci izens a e being in ec ed. While some o hese implemen a ions ha e been p o en o be e ec i e, o he ac o s such as inexpe ienced heal h wo ke s, poo heal h sys ems, lack o social dis ancing, and sca ce knowledge o he co ona i us make his p oblem all he mo e di icul o esol e. 1.1 Mo i a ion Va ious dashboa ds, bo h online and o line, ha e al eady been de eloped o analy ical use, wi h some ha ing been de eloped o isually ep esen he de- elopmen o he COVID-19 pandemic in nume ous aspec s. Va ious ea u es 8 CHAPTER 2. STATE OF THE ART 15 Figu e 2.1: ISCIII COVID-19 Dashboa d (Applica ion Used: Shiny) may o may no be de ailed. Fo example, a dashboa d ha uses da a s o ed in he cloud will ha e he capaci y o p esen da a in a deg ee o speci ici y simila o dashboa ds o na ional scope, hanks o he abili y o cloud-based pla o ms o ha e a i ually limi less amoun o s o age [9], while dashboa ds ha a e mo e limi ed in s o age may op o show summa ies o da a o each coun y ins ead. Co idly [10] (see Figu e 2.2) is an example o a dashboa d wi h a global scope, p ojec ing da a, such as cases, dea hs, eco e ies, and o he a iables ela ed o he pandemic, summa ized o each coun y. Hyb id Scope While he e a e dashboa ds ha ei he ha e a na ional scope o a global one, he e a e ce ainly dashboa ds ha a e a mix o he wo, ca ying bo h na ional da a o a ious coun ies and local da a o ce ain coun ies, as seen in he wo maps p esen ed by he COVID-19 Dashboa d by he Cen e o Sys ems Science and Enginee ing (CSSE) a Johns Hopkins Uni e si y (JHU) [11] ha show bo h he na ional da a o di e en coun ies and he da a o he coun ies in he Uni ed S a es o Ame ica (see Figu es 2.3 and 2.4). CHAPTER 2. STATE OF THE ART 16 Figu e 2.2: Co idly Dashboa d (Applica ion Used: Mapbox) As wi h dashboa ds o global scope, he speci ici y o he da a on a pe - coun y le el may depend on he capaci y o he da abase being used o he applica ion. Figu e 2.3: Global Map ound in he COVID-19 Dashboa d by JHU (Appli- ca ion Used: A cGIS Dashboa ds) CHAPTER 2. STATE OF THE ART 17 Figu e 2.4: U.S. Map ound in he COVID-19 Dashboa d by JHU (Applica- ion Used: A cGIS Dashboa ds) 2.1.2 Types o Dashboa ds by Type o Da a Used In e ms o he da a ha a e o en analyzed and p esen ed in exis ing dash- boa ds, i can be obse ed ha dashboa ds use one o he ou : epidemiolog- ical da a, da a ega ding clinical ials, da a ha desc ibe he se op e alence o he Co ona i us in a ce ain a ea, and da a on he p og ess o accine de elopmen . Epidemiological Da a T acke This is he ype o dashboa d ha is usually associa ed wi h he e m “COVID-19 dashboa d”. dashboa ds cen e ed a ound epidemiological da a ocus mos ly on he numbe o con i med cases and dea hs, shown bo h on a daily basis and as an accumula ion o e a ce ain pe iod o ime. O he da a ela ed o he de elopmen o he pandemic, including es ing a e, hospi al- iza ion coun , in ensi e ca e uni (ICU) admission coun , and eco e ies, may also be included and p esen ed o p o ide use s a mo e in-dep h analysis. As p e iously men ioned, Epidemiological Da a T acke s a e he mos common ype o COVID-19 dashboa ds. Hence, many well-known dash- boa ds, such as he Dashboa d o JHU and he Wo ld Heal h O ganiza ion (WHO) COVID-19 Dashboa d [12] (see Figu e 2.5) a e good examples. CHAPTER 2. STATE OF THE ART 18 Figu e 2.5: WHO COVID-19 Dashboa d (Applica ion Used: Sp inkl ) Clinical T ial T acke This ype o dashboa d aims o p esen use s wi h upda es ega ding he clinical ial esea ch o COVID-19 in e en ion me hods like accines, an- i i al d ugs, and e en adi ional Chinese medicine. Usual p esen ed da a include ial ID, ial loca ion, ial s a da e, ial comple ion da e (i ap- plicable), ype o ea men , p esen ed ou come and he URL o he esul s. Examples include he Co id-19 T ialsT acke [13] (see Figu e 2.6), a p od- uc o The Da aLab a he Uni e si y o Ox o d. I uses da a om he In e - na ional Clinical T ials Regis y Pla o m (ICTRP) o display clinical ials ela ed o he ea men o COVID-19 h ough ables and g aphs. Ano he example is Co id-T ials.o g [14, 15](see Figu e 2.7), whose unc ionali y is simila , wi h he excep ion o being able o display he da a in a map as well. Se op e alence Dashboa d Se op e alence Dashboa ds, such as Se oT acke [16] (see Figu e 2.8), o- cus on se op e alence o COVID-19 an ibodies. This means ha dashboa ds o his ype p esen he equency o indi iduals in a ce ain popula ion ha con ain COVID-19 an ibodies. CHAPTER 2. STATE OF THE ART 19 Figu e 2.6: Co id-19 T ialsT acke Dashboa d (Applica ion Used: Tableau) Figu e 2.7: Co id-T ials.o g Dashboa d (Applica ion Used: Cy el) Fo heal h au ho i ies, se op e alence may be an in e es ing a iable o obse e, as i plays a ole in an ibody es ing, a c ucial ac o in bo h moni o - ing he e olu ion o he Co ona i us and es ima ing he a ali y a e caused by in ec ions [17]. CHAPTER 2. STATE OF THE ART 20 Figu e 2.8: Se oT acke Dashboa d (Applica ions Used: Recha s and Lea le ) Vaccine T acke This ype o dashboa d acks he p og ess o he de elopmen o ac- cines agains he Co ona i us. The name o he accine and he cu en s a us/phase o he accine in he de elopmen p ocess is included. Addi- ionally, accines can be il e ed acco ding o hei cu en s a us. A good example o his is The Vaccine T acke [18] (see Figu e 2.9). Figu e 2.9: The Vaccine T acke (Applica ions Used: Ga sby and Reac ) CHAPTER 2. STATE OF THE ART 21 2.2 Applica ions o Dashboa d C ea ion and Con igu a ion The e a e nume ous applica ions a ailable on he ma ke ha specialize on he c ea ion and con igu a ion o dashboa ds, bo h paid and ee. Fo illus- a ion pu poses, six well-known applica ions a e discussed: Shiny, Powe BI, G a ana, Sisense, Domo, and A cGIS. Shiny Shiny [19] is an open-sou ce package based on R [20], a language and en i onmen commonly used o s a is ical p og amming ha allows an easy c ea ion and con igu a ion o in e ac i e web-based applica ions, including dashboa ds. Shiny applica ions can be made using RS udio [21], he in e- g a ed de elopmen en i onmen (IDE) o R, and hey can also be ex ended h ough CSS, h mlwidge s, and Ja aSc ip . A huge ad an age ha Shiny has o e o he applica ions o dashboa d c ea ion is ha i is ee o use, wi hou imposing limi a ions o any so on i s comple e unc ionali y. Any use can c ea e a ully unc ional dashboa d h ough Shiny wi hou incu ing any cos s. De eloped p oduc s can ei he be deployed o he Shinyapps.io cloud, wi h bo h ee and paid op ions [22], o be in eg a ed on-p emises wi h an exis ing se e . Powe BI Powe BI [23] is a business in elligence (BI) so wa e c ea ed by Mic oso . I allows da a modeling and isualiza ion h ough he use o a i icial in- elligence (AI). In addi ion, Powe BI p o ides in e ope abili y wi h o he Mic oso applica ions such as Azu e and Excel, end- o-end da a p o ec ion, and nume ous da a connec o s ha pe mi connec ions o da a sou ces such as Azu e SQL Da abase and Excel. While Powe BI is a paid se ice, an op ion o a ee ial is also a ailable o hose who wan o ge an idea o how he applica ion wo ks i s hand. G a ana G a ana [24] is a ee and open-sou ce pla o m ha ea u es suppo o o e 30 da abases, including G aphi e, In luxDB, and P ome heus, o CHAPTER 2. STATE OF THE ART 22 use in a single dashboa d and has a wide ange o dashboa ds and plugins a ailable in i s o icial lib a y. I also allows he de elopmen o dashboa ds ia collabo a ion and he de ini ion o ale s ( o example, when a ce ain h eshold alue is su passed o he da a being moni o ed) ha no i ies he use s. Sisense Sisense [25] is a cloud-na i e pla o m ha gi es i s use s he unc ions necessa y o c ea e powe ul analy ical solu ions such as embedded analy ics, da a mashups, access con ols a sys em, objec , da a, and p ocess le els, and cus omized ale s. I o e s P oduc s can be deployed ei he ia a p i a e cloud, on commodi y ha dwa e, o by aking a hyb id app oach. Sisense also o e s da a connec o s o AWS, Snow lake, and Google BigQue y. Being a BI pla o m ha ocuses on analy ics o en e p ise da a, his applica ion does no o e any ee e sions. Domo Domo [26] is a cloud-based BI pla o m ha consis s o h ee laye s: da a in eg a ion, business in elligence and analy ics, and he c ea ion o in elligen applica ions. I boas s o as pe o mance, unlimi ed scalabili y, da a au- oma ion, and a la ge a ay o da a connec o s, om da abases o e en social ne wo ks, allowing use s om businesses o so wa e de elope s o ocus on c ea ing he p oduc wi hou wo ying abou he unde lying in as uc u e. As wi h Powe BI, Domo also has a ee ial op ion o use s who wan o y i s capabili ies in connec ing, ans o ming, and isualizing da a. A cGIS Dashboa ds An online and a desk op pla o m c ea ed by Es i, A cGIS Dashboa ds [27] allows he c ea ion o dashboa ds ha ea u e loca ion-based analy ics as i s main ad an age o e o he pla o ms. I also has lexibili y, con igu abili y, a sui e o a ailable and eady- o-use da a isualiza ion ools like maps, lis s, and cha s, as well as ools ha allow use s o in e ac wi h dashboa ds c ea ed h ough his applica ion. This applica ion usually equi es o be pu chased o use, bu i also o e s a 21-day ee ial pe iod, gi en ha i is o non-p oduc ion use only. CHAPTER 2. STATE OF THE ART 23 2.3 Technical Requi emen s o C ea ing a Dash- boa d The ypical a chi ec u e o a dashboa d consis s o ou laye s, as shown in Figu e 2.10 [28]. As such, hese will make up he echnical equi emen s ha mus be complied in o de o c ea e a ully unc ioning dashboa d. Figu e 2.10: Typical a chi ec u e o a Dashboa d F on End Being he pa o he applica ion wi h which he use in e ac s he mos , i is undeniably he mos impo an pa o he dashboa d. The in e ace implemen ed in his laye mus be use - iendly and in e ac i e. Ca e ul layou planning and selec ion o g aphs and o he ypes o isual da a play a signi ican ole in helping he use s unde s and and in e p e clea ly he da a being shown. In addi ion, accessibili y ea u es, such as app op ia ely sized ex and labels and g aph anno a ions, mus also be conside ed. Back End The back end is whe e he se e pe o ms mos o i s ope a ions. I con ains a ious sc ip s and sou ce codes o pe o ming analyses and ans- o ma ions on he da a collec ed, as well as sc ip s o making p edic ions CHAPTER 2. STATE OF THE ART 24 and decisions based on he da a. In he case o dashboa ds, he main goal o his laye is o p epa e all da a o isualiza ion in he on end. Mechanisms o da a s o age, such as da abases used by he se e , can also be ound he e. As such, knowledge o basic da abase concep s such as c ea ion, inse ion, dele ion, agg ega ion, and p ojec ion a e a mus . I is also impo an o know he ad an ages and disad an ages o ela ional and non- ela ional da abases o de e mine which ype o da abase is bes sui ed o he applica ion. In he case o dashboa ds, ela ional da abases, many o which a e c ea ed using SQL, a e ideal as his makes da a analysis and p esen a ion easie o do. Da a Acquisi ion As explained by i s name, he da a acquisi ion laye consis s o sc ip s whose objec i e is o ob ain da a om sou ce sys ems and place ha da a as me ic alues in he s o age ound in he back end. The e a e mul iple me hods o acqui ing da a om hese sou ces, om s a ic analysis o da a mining. Hence, knowledge o hese p ocesses, as well as knowledge o he s uc u es ound in he sou ce sys ems, is a p e equisi e o his laye o unc ion p ope ly. Sou ce Sys ems While sou ce sys ems a e no ac ually pa o he dashboa d, hey s ill hold a c ucial pa in i s ope a ions. These sys ems s o e da a ha may be acqui ed o pe o m analyses, ans o ma ions, and isualiza ions. Ex- amples include code eposi o ies like Gi Hub and e en da abases om o he websi es. CHAPTER 3. DATA OF INTEREST FOR THE PROJECT 31 3.3 Va iables and Indica o s o In e es Since he beginning o he COVID-19 pandemic, a ious models ha e been de eloped in an a emp o accu a ely desc ibe and p edic he e olu ion o he i us. These models include di e en a iables, some o which we e disco e ed and calcula ed by means o scien i ic in es iga ions. As he model o be used o he dashboa d is inspi ed by he low ne wo k models p esen ed by López and Cukić [34, 35], as well as Kucha ski e al [36], he eam would use no only a iables commonly used in exis ing da abases o COVID-19 su eillance, bu also hose ha a e ela ed o he model and can be de i ed ma hema ically om he commonly ound a iables, such as accumula ed cases and dea hs, as well as indica o s such as he ansmission and dange indices and he accumula ed incidence, which will be discussed sho ly. 3.3.1 Va iables Daily Posi i e Cases This pe ains o he numbe o posi i e COVID-19 cases eco ded in a 24-hou in e al. Fu he mo e, his a iable can be classi ied acco ding o he he ype o diagnos ic es used o con i m he posi i e cases [37]: •An igen Tes s: Using samples collec ed h ough nasal o h oa swabs, an igen es s a e used o de ec a p o ein ha makes up pa o he co ona i us and a e use ul o iden i ying cases ha a e app oaching s a es o peak in ec ion. Compa ed o he o he ypes o es s, an igen es s a e cheape and as e , bu a e also less accu a e, being able o p oduce alse posi i es and alse nega i es. •Molecula /PCR Tes s: Also using nasal and h oa swabs, molec- ula es s ocus on iden i ying he gene ic makeup o he co ona i us using di e en me hods, wi h he mos amous one being he poly- me ase chain eac ion (PCR). While some molecula es s ha e shown alse nega i e esul s o up o 20% o he ime, his ype o es emains mo e accu a e in iden i ying posi i e cases han an igen es s. CHAPTER 3. DATA OF INTEREST FOR THE PROJECT 32 Dea hs These a e he numbe o dea hs caused by he co ona i us disease. We assume hese igu es a e epo ed by he hospi als daily. Reco e ies These a e pa ien s who ha e been discha ged om he hospi al o hose ha a e home-qua an ined and ha e been conside ed eco e ed by hei a - ending physicians. This a iable will emain use ul when aking in o accoun he o al numbe o eco e ies epo ed. Accumula ed Posi i e Cases As explained in i s name, his a iable ga he s he o al numbe o posi i e cases epo ed so a . This is calcula ed by e ie ing he sum o daily posi i e case coun s epo ed om s a o p esen . Ac i e Cases These a e posi i e cases ha a e nei he coun ed as a dea h no a eco e y, meaning ha hese a e ongoing cases o COVID-19. As a pe son in ec ed wi h COVID-19 may expe ience mild o se e e symp- oms, i is ecommended ha no all ac i e COVID-19 cases be conside ed as hospi aliza ions, bu a he be b oken down in o he ollowing ca ego ies: •Home-Qua an ined: Pa ien s who expe ience mild symp oms a e no ecommended o be hospi alized ( o allow hospi als o accommoda e cases o highe se e i y), so hey a e qua an ined a home ins ead. They a e usually moni o ed by hei espec i e physicians h ough emo e means such as examina ions o e he phone. •Hospi alized: Pa ien s who expe ience se e e symp oms and a e in need o hospi al ca e all unde his ca ego y. These pa ien s s ay in his s a e un il hey eco e and a e deemed sa e o discha ge o un il hei condi ion wo sens in o a c i ical s a e. •C i ical: C i ical cases in ol e pa ien s who a e expe iencing ex eme symp oms and equi e immedia e a en ion by being admi ed in o in- ensi e ca e uni s (ICUs). CHAPTER 3. DATA OF INTEREST FOR THE PROJECT 33 Numbe o Tes s Done This consis s o he accumula ed numbe o es s pe o med in a ce ain a ea. This will include bo h es s ha e u ned posi i e esul s as well as nega i e ones. To p oduce mo e comp ehensible numbe s, his numbe may be p esen ed as a a io o es s pe 1M/100,000 inhabi an s. Numbe o Cases pe 1M/100,000 Inhabi an s Compa ing he numbe o cases (be i ac i e cases, numbe o dea hs, o numbe o eco e ies) may be di icul o unde s and o he a e age use . This a iable aims o p ojec hese numbe s as a a io o numbe o cases o numbe o inhabi an s, p o iding a densi y o cases wi hin a pa o he popula ion. Popula ion-Weigh ed Densi y (PWD) PWD is one o he a iables ha explain he sp ead o he COVID-19 pandemic [38]. As his a iable desc ibes he densi y a which an a e age pe son li es, i can be obse ed ha his in luences he a e o dea hs caused by he co ona i us. Inc eased social dis ancing esul s in a educ ion in PWD (albei empo a ily), indica ing e ec i eness in measu es aken by a uni o go e nmen . 3.3.2 Indica o s T ansmission Index (R0) The ansmission index [34] ep esen s he powe o i us ansmission and can be in e ed by calcula ing he a io be ween he numbe o in ec ions eco ded a a ce ain day and he numbe o in ec ions eco ded 14 days be o e , as seen in Equa ion 3.2. A alue o 1 o g ea e indica es dange in he ansmission o he i us while a alue o less han 1 indica es no dange . c R0( ) = In ( ) In ( −14) (3.2) CHAPTER 3. DATA OF INTEREST FOR THE PROJECT 34 Cumula i e Incidence pe 100,000 Inhabi an s (Inc) As seen in Equa ion 3.3, his index de e mines he cumula i e numbe o posi i e COVID-19 cases in he pas 14 days pe 100,000 inhabi an s [34, 35]. As his a iable akes in o conside a ion he popula ion o he coun y, i is use ul o making ai e compa isons be ween he si ua ions o di e en coun ies wi h ega ds o he pandemic. Howe e , i s disad an age comes om he ac ha his indica o only conside s he numbe o in ec ed people. Inci= ( 14 X j=1 Ci−j)∗100000/P, ∀i≥15 (3.3) Dange Index (DI) The dange index [34] is used o de e mine he le el o dange a low ne wo k p esen s. Using Equa ion 3.4, he DI ob ained indica es li le o no dange i i s alue is equal o o less han 0, while a highe DI indica es a mo e dange ous low ne wo k. While bo h R0 and Inc only conside s he numbe o in ec ed cases, DI has an ad an age such ha i conside s he numbe o posi i e cases, he numbe o dea hs, and he numbe o eco e ies, each o which can be weighed o p o ide mo e adequa e alues depending on he si ua ion. Howe e , com- pa ed o he Inc indica o , he popula ion o he coun y is no aken in o conside a ion. DI( ) = In ( ) + F( )−Rec( ) = X x∈V x,INF ( ) + X x∈V x,F ( )−X x∈V x,R( ) (3.4) CHAPTER 3. DATA OF INTEREST FOR THE PROJECT 35 3.3.3 Smoo hed Va iables and Indica o s In ob aining da a on COVID-19, he e a e imes when la ge peaks a e ol- lowed by sudden d ops in alue, possibly esul ing in o noisy da a [39]. To coun e his, da a smoo hing is applied by using an algo i hm o emo e noise om a da ase and allow pa e ns o s and ou mo e clea ly. In his case, he a iables and indica o s men ioned p e iously we e smoo hed by calcula ing he 7-day mo ing a e ages, ound in Equa ion 3.5. ∀i≥4, mm(xi) = 1 7 i+3 X k=i−3 (xi)(3.5) Chap e 4 Da a P epa a ion A e ha ing a lis o possible sou ces o da a o he p ojec , he da a would need o be p epa ed o use in he dashboa d. This p epa a ion p ocess includes ope a ions such as uni ying he o ma o da a, cleaning ou in alid da a, and placing he esul s oge he in a ile ha he dashboa d can access and use o p esen he da a isually. In his pa o he p ojec , a collabo a ion was ini ia ed wi h ano he g oup, whose goal is o sea ch and analyze a ailable in o ma ion on he COVID-19 pandemic. They would ul ima ely be in cha ge o p o iding he da ase o be used by he dashboa d. 4.1 Da abases Used as Bases o Da abase C e- a ion As p e iously explained, he pa ne g oup was in cha ge o ob aining a ail- able in o ma ion on he si ua ions o each coun y. While mos in o ma ion can be ound on he websi e o each coun y’s heal h agency (which was he case o majo coun ies wi h conside ably eliably go e nmen s such as Spain, Japan, and Russia), he ollowing eposi o ies a e highligh ed due o he ac ha hey con ain in o ma ion ha has al eady been compiled and is cons an ly upda ed o con enience in da a ga he ing: •h ps://gi hub.com/da ase s/co id-19 •h ps://gi hub.com/cssegisandda a/co id-19 36 CHAPTER 4. DATA PREPARATION 37 4.2 Da ase C ea ion o P ojec Use 4.2.1 Fi s Da ase In o de o ini ialize he he de elopmen o he p ojec , ano he da ase mus be empo a ily used o es ing he unc ionali y o he dashboa d. Fo his pu pose, a da ase p o ided by he p ojec di ec o s ha con ained da a on he pandemic si ua ion in 14 coun ies was used. The ollowing columns a e ound in his da ase : •COUNTRY: coun y in which he da a was eco ded •FECHA: da e o eco ding •CONTAGIADOS: daily numbe o posi i e cases o COVID-19 •FALLECIDOS: daily numbe o dea hs caused by COVID-19 •HOSPITALIZADOS: daily numbe o hospi aliza ions due o COVID-19 •UCIs: daily numbe o COVID-19 pa ien s admi ed in o ICUs 4.2.2 Second Da ase C ea ed by he pa ne g oup, his da ase is sepa a ed in o wo iles: one con aining da a pe aining o he Uni ed S a es alone and ano he con aining da a o he es o he wo ld. Compa ed o he p e ious da ase , his con- ains in o ma ion on mo e coun ies, as well as mo e columns ha conside a iables ha we e p e iously no p esen , such as he numbe o ac i e cases and he dange index [34, 35]. Ano he di e ence is ha , when applicable, da a is also o ganized by he p o inces o he in ol ed coun ies. This da ase consis s o he ollowing columns: •Coun y_Region: coun y/ egion in which he da a was eco ded •P o ince: p o ince o he coun y in which he da a was eco ded, i a ailable •La : la i udinal coo dina e o he a ea’s loca ion •Long: longi udinal coo dina e o he a ea’s loca ion CHAPTER 4. DATA PREPARATION 38 •Las _Upda e1: da e o eco ding •Con i med: cumula i e numbe o con i med posi i e COVID-19 cases •Dea hs: cumula i e numbe o dea hs caused by COVID-19 •Reco e ed: cumula i e numbe o COVID-19 eco e ies •Ac i e: cumula i e numbe o ac i e COVID-19 cases •Daily_Con i med: daily numbe o con i med posi i e COVID-19 cases •Daily_Dea hs: daily numbe o dea hs caused by COVID-19 •Daily_Reco e ed: daily numbe o COVID-19 eco e ies •Daily_Ac i e: daily numbe o ac i e COVID-19 cases •IP: dange index •R0: ansmission index •Daily_Con i medMM: mo ing a e age o he con i med posi i e COVID- 19 cases •Daily_Mue osMM: mo ing a e age o he dea hs cause by COVID-19 •Daily_Recupe adosMM: mo ing a e age o he COVID-19 eco e ies •Daily_Ac i osMM: mo ing a e age o he ac i e COVID-19 cases •IPMM: mo ing a e age o he dange index •R0MM: mo ing a e age o he ansmission index Meanwhile, he da ase pe aining o he Uni ed S a es con ained addi- ional columns: •Daily_Tes ed: daily numbe o people who ook he COVID-19 es •Daily_Hospi alized: daily numbe o people admi ed o he hospi al due o COVID-19 •People_Tes ed: cumula i e numbe o people who ook he COVID-19 es CHAPTER 4. DATA PREPARATION 39 •People_Hospi alized: cumula i e numbe o people admi ed o he hospi al due o COVID-19 4.2.3 Final Da ase Upon s udying he p e ious da ase , he ollowing issues we e encoun e ed: •The da ase con ained columns deemed necessa y o he dashboa d implemen a ion (i.e. La , Long). •The iles con aining he da ase has a o al size equal o 50MB. This was a po en ial p oblem wi h ega ds o he limi in he size o he applica ion o be uploaded in he se e . •Some o he da a con ain names o coun ies ha do no ma ch wi h hose used in he dashboa d, pa icula ly i s map unc ionali y. •The INC indica o , which was a a iable o in e es , was no included. •The calcula ion o he mo ing a e age was done by egion, his was a p oblem as he dashboa d being implemen ed was o ganized by coun y. The eam had c ea ed an R sc ip named da aT ans o m.R o sol e hese issues. This sc ip pa ses he da a om bo h da ase iles and pe o ms he necessa y ans o ma ions (i.e. column enaming, da a me ging, calcula ion o ex a a iables and smoo hing a iables and indica o s using mo ing a - e ages), c ea ing he inal da ase , da ase CODA.cs . The s uc u e o his da ase is explained in he Chap e 6. 4.3 Da ase Quali y E alua ion C ea ing he da ase o he dashboa d in ol ed mul iple asks ha equi ed no only ime and e o , bu also he equi ed knowledge o pe o m such asks and in o de o edac he da ase e icien ly. No only does he da ase ha e o be well-de ined, i also has o be eliable. One impo an ask in his p ocess is he e alua ion o he quali y o hese da ase s. I gi es he idea o how alid he in o ma ion inside a da ase can be. This can be es ima ed by some quali a i e ac o s p esen in he da a. CHAPTER 4. DATA PREPARATION 40 An impo an aspec abou dashboa ds is ha pa o hei quali y and e ec i eness is de i ed om he da ase i uses o p esen in o ma ion. As such, i is impo an o e alua e a leas he inal da ase used. Wo king wi h his da ase , he ollowing obse a ions, some o which led o modi ica ions in he p ojec , we e made: •Some coun ies do no ha e upda ed da a. •Some alues, like eco dings o nega i e dea hs, do no make sense and can be a ibu ed o e o s in eco ding da a. •Mos coun ies con ain ou lie da a due o inconsis encies such as in- co po a ing da a eco ded in a ce ain day in o ha o a la e da e. •Da a om coun ies such as No h Ko ea, An a c ica, Wes e n Saha a and Tu kmenis an a e una ailable and he e o e could no be shown in he dashboa d. CHAPTER 5. PROJECT DEVELOPMENT 47 Figu e 5.5: Final p o o ype design. Map View 3 Figu e 5.6: Final p o o ype design. G aph View CHAPTER 5. PROJECT DEVELOPMENT 48 Imp o emen s o Dashboa d Implemen a ion As ag eed in he p o o ype design, he dashboa d was spli in o wo sc eens/ abs, one o p esen ing use s he g aphs ha show he e olu ion o he di e en a iables in a ious coun ies (see Figu e 5.7) and ano he one o displaying he cumula i e numbe s o each coun y in a map (see Figu e 5.8). Addi ion- ally, he g aph unc ion allows use s o see he e olu ion o he pandemic in mul iple coun ies o mul iple a iables o allow compa isons. Figu e 5.7: Sp in 1 esul (g aph iew) Figu e 5.8: Sp in 1 esul (map iew) CHAPTER 5. PROJECT DEVELOPMENT 49 Sp in E alua ion A he end o he sp in , some o he asks included in he backlog o his i e a ion we e no implemen ed, such as he upda ing Lis View UI in he Map View. Ins ead, only an In o ma ion Panel UI ha con ained no da a was displayed. A mee ing was hen held wi h he p ojec di ec o o p o ide upda es on he p og ess o he p ojec , as well as p esen he dashboa d in i s s a e a e he sp in . Feedback was ecei ed om he di ec o , which would hen be aken in o conside a ion upon beginning he nex sp in . Some o hese included changing he names o he coun ies in he map iew o hei English names, as hey we e o iginally p esen ed as hei espec i e names in hei own languages, selec ing mo e han jus wo coun ies o a iables, including a legend o g aphs, and p esen ing a summa y o all da a in he in o ma ion panel by means o p esen ing, o example, he coun ies wi h he mos accumula ed numbe o posi i e cases. 5.4 Second Sp in Sp in Planning A e conduc ing a mee ing a he end o he i s sp in , we ha e selec ed asks o be included in he second sp in backlog as shown in Table 5.3. Also lea ning om ou expe ience in he i s sp in , some o he asks ha e been sligh ly modi ied o accommoda e he limi a ions p esen ed by Shiny, such as di icul ies in laye ing componen s o he dashboa d UI. As p e iously done, he asks ha e been di ided in o smalle ones o ensu e manageabili y by he eam membe s. Design As p e iously s a ed, he second sp in ocuses on adjus ing he imple- men a ion o he dashboa d o he design limi a ions o Shiny and on e ining he dashboa d con en s as much as possible. As such, he design p o o ypes ha e been e ined, as seen in Figu es 5.9 and 5.10. CHAPTER 5. PROJECT DEVELOPMENT 50 S o y Task 1 A. Implemen mouse ho e unc ion o show da a a a speci ic poin in he g aph B. Implemen legends in he g aph C. Show he axis labels bigge D. Fix da e axis scale o 1 week E. Fix da e axis angle o 45 deg ees F. Add con ols o isualize g aph (zoom in/ou , pan, au o-scale) G. Reloca e he panel o il e s abo e he g aph. 2 A. Se colo di ision o pe cen ages depending on he highes numbe o accumula ed cases o he chosen a iable B. Implemen mouse ho e unc ion o show he coun y’s name and he alue o he a iable chosen C. Change he map o show all coun ies in English 8 A. Limi o 5 coun ies o choose B. Remo e Single Va iable/Coun y Mode Radio Bu on and se de aul mode o Mul iple Coun y 9 A. Limi o 5 a iables o choose B. Es ablish a pale e o colo s o be used o e e y di e en line o be d awn in he g aph. 10 A. Change UI o slide o bo h Map and G aph Views B. Pu con ols o modi y bo h he s a and end da e in he G aph View 11 A. In es iga e o he g aphs ha could well desc ibe some a iables B. Implemen a adio bu on UI o choose be ween some ype o g aphs C. Implemen se e unc ions o isualize selec ed g aph 12 A. Implemen a able showing 5 coun ies wi h he highes numbe o accumula ed cases o he chosen a iable (o he i s a iable chosen in mul iple a iables mode) in he G aph View B. Implemen he same able in he Map View C. Implemen ano he In o ma ion Panel UI in he G aph View ha shows he o al numbe o in ec ed and dea h cases D. Implemen se e unc ions o calcula e hese a iables E. Reloca e he In o ma ion Panel and Fil e Selec Box UI in he Map View and place i o e he map Table 5.3: Second sp in backlog CHAPTER 5. PROJECT DEVELOPMENT 51 Figu e 5.9: Re ised p o o ype design (g aph iew) Figu e 5.10: Re ised p o o ype design (map iew) CHAPTER 5. PROJECT DEVELOPMENT 52 Imp o emen s o Dashboa d Implemen a ion One o he bigges changes in oduced in he dashboa d implemen a ion o his sp in is he op ion o selec mo e han 2 coun ies when selec ing he "mul iple coun ies" op ion. Use s would now be allowed o choose up o 5 coun ies o be displayed in he g aph. This limi was placed o allow use s o compa e di e en coun ies wi hou possibly o e c owding he esul ing g aph wi h a ious lines ha may become un eadable. Ano he majo change in he g aph iew o he dashboa d is he inclusion o box plo s. As discussed in Sec ion 3.2, box plo s a e used o de e mine he de ia ion o he da a and he nume ical ange hey occupy, which use s may ind in e es ing. The inal majo change in he g aph iew o he dashboa d is ha he in o ma ion panel om he p e ious sp in was eplaced wi h wo smalle in o ma ion panels ha p esen summa ies abou he cu en pandemic si u- a ion globally. Fo example, use s could o see he o al numbe o cases and dea hs eco ded wo ldwide. They could also ind a lis o he op coun ies in e ms o numbe o posi i e COVID-19 cases. Figu e 5.11: Sp in 2 esul (line g aph) Wi h espec o he map iew o he dashboa d, he coun y names we e all enamed o hei English names. This was done by eplacing he map CHAPTER 5. PROJECT DEVELOPMENT 53 Figu e 5.12: Sp in 2 esul (box plo ) used wi h he one by Es i, which also changed o he colo pale e o shades o blue. A legend was also placed o indica e he alues a ce ain colo in he map ep esen s. The panel o il e s oge he wi h he In o ma ion Panel we e placed o e he map o maximize he space o be occupied by he map i sel . Finally, he in o ma ion panel, implemen ed in he 1s i e a ion and ha con ained da a, is epu posed o p esen he coun ies wi h he highes numbe o COVID-19 in ec ions. Sp in E alua ion Feedback om he p ojec di ec o included changing he colo s om shades o blue o shades o ed, as his would appea mo e ala ming and would gi e he imp ession ha he e is an ongoing h ea , he inclusion o digi g oup sepa a o s in he numbe s p esen ed in he dashboa d o imp o e eadabili y, and changing he labels in he in o ma ion panels o he G aph View (i.e. changing "To al cases" o "Global numbe o cases") o p e en con usion. Addi ionally, i would p esen a bigge con enience o use s o he dashboa d o ha e a unc ion ha cap u es he g aph being shown, be i he en i e g aph o a po ion o i . CHAPTER 5. PROJECT DEVELOPMENT 54 Figu e 5.13: Sp in 2 esul (map iew) CHAPTER 5. PROJECT DEVELOPMENT 55 5.5 Thi d Sp in Sp in Planning The sp in began by selec ing he asks o be execu ed, elabo a ed in he sp in backlog ound in Table 5.4 S o y Task 1 A. P ocess he new da ase o adap i wi h he unc ions o he dashboa d B. Add he accumula ed a iables in he selec box UI C. Include Dange Index o he lis o a iables ha can be shown as g aphs 2A. Change he colo pale e o he map o indica e in ensi y o numbe s by shades o ed 5 A. Implemen UI o iew Coun y Table in he Map View B. Implemen Upda e Coun y Table UI se e unc ion C. Implemen map click unc ion o enable his ea u e 12 A. Imp o e he o ma ing o numbe s by placing digi g oup sepa a o s Table 5.4: Thi d sp in backlog Design Imp o ing on he dashboa d designs c ea ed in he p e ious sp in and conside ing he eedback ecei ed in he p e ious sp in e alua ion, he p e- sen a ion o he da a and he labels ound in he g aph iew o he dashboa d we e enhanced (see Figu e 5.14). The colo s used in he map we e changed o shades o ed and a panel ha upda es whene e he use clicks on a coun y in he map o see mo e in o ma ion was included (see Figu e 5.15). CHAPTER 5. PROJECT DEVELOPMENT 56 Figu e 5.14: Re ised p o o ype design (g aph iew) Figu e 5.15: Re ised p o o ype design (map iew) CHAPTER 5. PROJECT DEVELOPMENT 63 Figu e 5.20: Websi e design (home iew) CHAPTER 5. PROJECT DEVELOPMENT 64 Imp o emen s o Dashboa d Implemen a ion The mos signi ican implemen a ion in his sp in is he c ea ion o he websi e in Google Si es. In he websi e, bo h he comple ely de eloped dash- boa d and addi ional in o ma ion on bo h he p ojec i sel and he eam ha made he dashboa d we e placed. By c ea ing and using a CSS ile s yle.css o indica e he s yle adjus men s and desc ibe how HTML ele- men s will be displayed in he web page (ma gins, paddings, e c.), he design o he websi e was made such ha i allowed he dashboa d o i wi hin he alloca ed space when in eg a ed. The CSS ile is included along wi h he sou ce code. Ano he impo an change is ha he sou ce code was e ac o ed, esul - ing in a mo e e icien ou pu o da a by he dashboa d. The memo y usage o he dashboa d is also kep below 1 GB, allowing he eam o con inue using he ee plan o Shinyapps.io. O he changes made in his i e a ion include he addi ion o he Dange Index, R0 and Inc a iables in he g aph iew o he dashboa d. The alues unde he Dange Index, R0, Inc columns we e calcula ed using Equa ions 3.5, 3.2, 3.3 ound in Chap e 3. In addi ion, he Wo ld Popula ions P ospec s 2019 da ase published by he Uni ed Na ions [42] was u ilized o ob ain he popula ions o each coun y, which we e used o calcula e Inc. Some mino design changes, such as he swapping o o al and pa ial ables in he map iew o he dashboa d and he addi ion o da a poin s in he plo o allow u he isualiza ion o he g aph, we e also made. CHAPTER 5. PROJECT DEVELOPMENT 65 Figu e 5.21: Websi e implemen a ion (home iew) Figu e 5.22: Websi e implemen a ion(abou us iew) CHAPTER 5. PROJECT DEVELOPMENT 66 Figu e 5.23: Websi e implemen a ion (abou his p ojec iew) CHAPTER 5. PROJECT DEVELOPMENT 67 Sp in E alua ion In his sp in e alua ion, he web page’s design was well- ecei ed. The la es de elopmen s done in he dashboa d’s implemen a ion and i s design we e all accep ed. Mino adjus men s, like co ec ing g amma ical e o s in labels in he dashboa d and changing he adio bu ons o a d op-down lis o he a i- ables in he map, would ha e o be made. Some addi ional in o ma ion, such as he use manual, we e also equi ed in he Abou This P ojec page o he websi e. These mino co ec ions we e hen made immedia ely and we e included in he sp in backlog o his i e a ion. Wi h all c i e ia lis ed in he P oduc Backlog ul illed and e alua ed, he de elopmen phase o he dashboa d was concluded and deployed o assessmen . 5.7 RS udio Lib a ies Used Aside om he base RS udio lib a y, which is used whene e he RS udio IDE is launched, he ollowing lib a ies we e u ilized in he de elopmen o he dashboa d: •shiny: R amewo k used o build he applica ion’s UI and se e in e ac ions •shinydashboa d: used o design he s uc u e o he panels o he applica ion •dply : used o pa sing, p ocessing and c ea ing da a ames in he Rs udio IDE •lea le : used o c ea ing and isualizing he map •ggplo 2: used o c ea ing and isualizing he g aph • gdal: used o pa sing a shape ile which is hen needed o c ea e and d aw polygon laye s on he map • eshape2: used o me ging da a ames mo e e ec i ely han dply CHAPTER 5. PROJECT DEVELOPMENT 68 •h ml ools: used o manipula ing HTML elemen s such as he ho e ool on he map •plo ly: used o he addi ional ea u es o he g aph such as he oolba and he ho e unc ion In summa y, he dashboa d con ains he a chi ec u e seen in Figu e 5.24. Figu e 5.24: Dashboa d a chi ec u e 5.8 Sc um Sp in Schedule Figu e 5.25 p esen s he da es in which each sp in was execu ed. The du a- ion o each sp in was app oxima ely wo o h ee weeks. Al hough he i s sp in , ha las ed om Janua y 11, 2021 o Feb ua y 19, 2021, seemed ex ensi e, he e ec i e de elopmen ime on he p ojec was 2 weeks, as his pe iod coincided wi h he comple ion o nume ous academic equi emen s. CHAPTER 5. PROJECT DEVELOPMENT 69 Figu e 5.25: Realized sp in da es Chap e 6 Resul s In his chap e , he esul s o he de elopmen o he dashboa d and he co esponding web page a e p esen ed. To p ope ly in e p e hese esul s, some assump ions mus be made. One assump ion is ha he da a p o ided by each coun y is upda ed almos daily. Ano he assump ion is ha he da a u h ully e lec s he si ua ion o he pandemic o each coun y. The inal assump ion is ha he alues eco ded o each day do no con ain e o s ha d as ically a ec p e ious da a. Howe e , in some pe iods o ime, some coun ies ha e p o ided da a ha iola e hese assump ions, as hey we e published in such manne . In he i s sec ion, he inal da ase is explained, along wi h i s s uc u e. The second sec ion p esen s he comple ed dashboa d and websi e, along wi h a sho guide ha explains hei usage. To u he demons a e he use o he dashboa d and he websi e, he hi d sec ion p esen s an example analysis ha compa es he si ua ion o he pandemic in Spain and in he Philippines. The inal sec ion o his chap e p o ides a guide on in eg a ing he dashboa d o o he websi es. 6.1 Final Da ase The inal da ase c ea ed o he use o he dashboa d was also made a ailable in a public Google D i e olde , ound a : •h ps:// inyu l.com/CODAda ase . This da ase con ains he ollowing columns, each o which may be used o u he analyses: 70 CHAPTER 6. RESULTS 71 •COUNTRY: coun y/ egion in which he da a was eco ded •DATE: da e o eco ding •DAILY_CONFIRMED: daily numbe o con i med posi i e COVID-19 cases •DAILY_DEATHS: daily numbe o dea hs caused by COVID-19 •DAILY_RECOVERED: daily numbe o COVID-19 eco e ies •ACTIVE: cumula i e numbe o ac i e COVID-19 cases •CUMULATIVE_CONFIRMED: cumula i e numbe o con i med posi i e COVID- 19 cases •CUMULATIVE_DEATHS: cumula i e numbe o dea hs caused by COVID- 19 •CUMULATIVE_RECOVERED: cumula i e numbe o COVID-19 eco e ies •DANGER_INDEX: dange index as o he da e o eco ding •R0: ansmission index as o he da e o eco ding •INC: cumula i e incidence pe 100,000 habi an s as o he da e o eco d- ing •SMOOTH_*: smoo hed a iable/indica o using mo ing a e age Addi ionally, he Inc alues o Spain in his da ase (see Figu e 6.1) was compa ed o he g aph c ea ed using he o icial published da a o he Spanish Minis y o Heal h [43] (see Figu e 6.2) o e i y he accu acy o he calcu- la ions done o ob ain he indica o . The wo se s o da a co e he same ange o da es, om July 2020 o May 2021. As a esul , he alues ound in he dashboa d a e sligh ly highe han hose ound in he o icial da a, which may be a ibu ed o sligh ly di e en alues used o he popula ion o Spain. None heless, he pa e n c ea ed by he g aph in he dashboa d ma ches ha in he o icial g aph, meaning ha he calcula ions we e success ul. CHAPTER 6. RESULTS 72 Figu e 6.1: Inc o Spain om July 2020 o May 2021, as shown in Dashboa d Figu e 6.2: Inc o Spain om July 2020 o May 2021 6.2 Dashboa d and Websi e The websi e c ea ed h oughou he du a ion o his p ojec (see Figu e 6.3) was published online. The ollowing links may be accessed o ind he dashboa d and websi e, as well as he sou ce code o he p ojec : CHAPTER 6. RESULTS 79 Figu e 6.10: Map compa ing Spain and he Philippines unde Inc on June 8, 2020 6.4 In eg a ing he Dashboa d o O he Web- si es As he de eloped dashboa d, being deployed in Shinyapps.io, is cloud-based, use s can in eg a e he dashboa d in o hei own websi es by simply adding i as an HTML elemen using he ollowing code: <i ame heigh ="720" wid h="100%" amebo de ="no" s c="h ps://co id-2019-dashboa d.shinyapps.io/dashboa d" /> The ad an age o his is ha , using he heigh and wid h a ibu es, use s may also change he dimensions o he dashboa d o help i i in any websi e layou . Chap e 7 Indi idual Con ibu ions Th oughou he en i e de elopmen o he p ojec , each membe has con- ibu ed in e e y phase, be i h ough esea ch, h ough implemen a ion, o h ough e i ica ion. Fo ull anspa ency, each eam membe has ex- pounded his own con ibu ions in his chap e . 7.1 Con ibu ions made by Ad ián Ga cía Sánchez A e he es ablishmen o he main heme o he p ojec , he objec i es and mo i a ion, I was in cha ge o es ablishing he wo k plan oge he wi h he s uc u e o he documen , which has allowed us du ing he whole p ojec o be able o be o ganized and ocused on ou objec i es, hus a oiding some isks ela ed o he non-achie emen o he objec i es o o he diso ganiza- ion and was e o ime and e o . In he esea ch phase I ocused on he echnical equi emen s o he c ea ion o a dashboa d, such as he di e en laye s ha e e y unc ional dashboa d equi es. These a e he backend, which is he pa in cha ge o making all he logic o a page wo k, he on end, which is he pa o he so wa e ha in e ac s wi h he use s, and o cou se he da a om which he dashboa d is ed. Toge he wi h Ha old I was in esponsible o esea ching abou he exis ing da abases on he in e ne ela ed o his opic and he mos use ul ypes o g aphics o his p ojec , such as line g aphs, maps o box plo s among o he s. All his allowed us o ge an idea o he main poin s o a good dashboa d and hose hings we should a oid in ou p ojec . In he c ea ion o he da ase , I helped in he e alua ion o he da ase s, 80 CHAPTER 7. INDIVIDUAL CONTRIBUTIONS 81 o ind possible isks in hei c ea ion o ela ed o duplica es o a bad use o space, since shiny has a 1GB limi o i s ee plan, which is he one we use. I did his by examining he con en s o he da ase ho oughly using R and RS udio. A e he esea ch phase, he h ee o us c ea ed he planning o ollow, bo h he ools o use (R s udio, Shiny, Google Si es) and he me hodology o use (Sc um) and also he du a ion o each i e a ion and he di ision o asks. I was in cha ge o dis ibu ing and planning each o he i e a ions, he p oduc backlog ha includes he assignmen o p io i ies o he asks, based on he ini ial planning and on he p oblems ha a ose in he de elopmen ha made us adap he planning om ime o ime. We chose Sc um o e o he agile me hodologies o ha e his lexibili y when changing he planning. I also did he dashboa d design o each i e a ion, based on he limi a ions encoun e ed in he de elopmen and abo e all o make an in ui i e and simple ool o all he public. In he de elopmen phase, I helped in implemen ing he map by s udying he lea le lib a y and he laye s ha can be included wi h he map such as he legends and he polygons and shapes he de ine each coun y. I also pe o med es ing and debugging asks o ind he bugs and hings o imp o e in he sou ce code. A e each o he i e a ions we me wi h he di ec o and es ablished he conclusions and p epa ed he basis o he nex i e a ion. Rega ding he websi e I ook ca e o he c ea ion and he design and o ganiza ion o he pages, aking in o accoun he many limi a ions ha google si es has, bu enough o ou pu pose. The si e was o ganized ocusing on he dashboa d, which is he mos impo an piece and occupies mos o he main page. We added se e al ex a pages explaining a ious aspec s o he wo k, such as i s use manual o da a eposi o ies. We also ound ways o con ac he membe s o he g oup. I also poin ed ou he possible modi ica ions and adjus men s conce ning he aes he ics such as he on sizes, he posi ioning o he panels, he colo s, he ma gins, e c. This is o imp o e he isualiza ion o he dashboa d in he web page. Finally, we es ablished oge he he conclusions, whe e we commen ed all he knowledge acqui ed du ing he elabo a ion o he p ojec , bo h a he le el o ools and da a analysis o me hodology. Then we de ined some u u e wo k, whe e we es ablish possible lines o ac ion o he con inua ion o he p ojec . I was also in cha ge o he modi ica ion o he epo , in la ex and h ough he O e lea pla o m, adjus ing he o ma and ixing he e o s. CHAPTER 7. INDIVIDUAL CONTRIBUTIONS 82 7.2 Con ibu ions made by Dale F ancis Valen- cia Calicdan A he s a o he p ojec , i s hing we did was speci y he main opic o his p ojec . We de ined he subjec his applica ion was abou which was he co ona i us, i s de ini ion, o igin, and he impo ance o s udying i s e olu ion. Wi h ha , I de ined he objec i es his whole p ojec will enci cle in o and he main goals he eam should achie e by he end o his p ojec ’s du a ion. As we go on in o he in es iga ion phase o he p ojec , I was in cha ge o s udying he di e en a iables in ol ed in s udying he e olu ion o he co ona i us disease. I spen ime eading a icles and looking h ough he a ious dashboa ds my colleagues we e in es iga ing o iden i y he common a iables and o he s ha we e signi ican . I e alua ed hem by which I chose he mos impo an ones which could be o in e es in de eloping he dashboa d. Upon he end o he in es iga ion phase, we mo ed on in o he de elop- men phase. In his phase, we i s laid ou he plan he eam should ollow, he me hodology o use, which was Sc um me hodology, he du a ion o each i e a ion, and he pa i ion o asks. In each i e a ion I was p ima ily ocused in building he sou ce code o he p og am in R language, while s ic ly ol- lowing he sp in backlog and he design es ablished o e e y i e a ion. This also included in esea ching and s udying a ious lib a ies which could be o use in de eloping he dashboa d. By he end o e e y i e a ion, I also epo ed he a ious limi a ions I ha e encoun e ed. This led in o sugges- ions on how o imp o e he dashboa d o he nex i e a ion. By he las i e a ion, I uploaded he dashboa d in o he dedica ed se e which included signing up an accoun , uploading he sou ce code and esou ces, and inally, deploying he applica ion. A he begginning o his phase, we wo ked on an ini ial da abase p o ided by he di ec o s o he p ojec , which is a small da ase con aining eco dings o co ona i us cases o some coun ies. By he 3 d i e a ion, he i s e sion o he p oposed da ase , c ea ed by he o he g oup, was sen o us. The o ma o his da ase (numbe o columns, column names, a iables a ailable, e c.) was di e en om he one we p e iously wo ked on. I es uc u ed he code and adap ed i o he o ma he p oposed da ase used. La e on i was seen ha he da a in he da ase p o ided by he pa ne CHAPTER 7. INDIVIDUAL CONTRIBUTIONS 83 g oup was o ganized by egions while he dashboa d is s uc u ed o show da a by coun y. We also lea ned ha calcula ions done inside he sou ce code akes a lo o memo y. Wi h hese indings, he eam decided o c ea e ano he da ase by ans o ming he ones we we e p o ided. I c ea ed he sc ip which pe o ms he necessa y ans o ma ion o he da ase , as well as he calcula ions o some a iables, o a oid unnecessa y ope a ions while he applica ion is deployed in he se e . By he end o he de elopmen phase, my colleagues we e implemen ing he web page o he dashboa d. A e discussing some modi ica ions and possible adjus men s o he HTML elemen s o he dashboa d, o i o i in he alloca ed window in he web page, I c ea ed a CSS ile and adjus ed he ma gins and paddings o he dashboa d as well as some o i s aes he ics o u he imp o e i s accessibili y and usabili y. A e he de elopmen phase, I con ibu ed in edac ing he documen a- ion o his p ojec , assis ing in he chap e s I was p ima ily ocused a . I explained he inal da ase ha was one o he esul s o his p ojec , and p oo ead he en i e pa o Chap e 5 explaining he p ojec ’s de elopmen . CHAPTER 7. INDIVIDUAL CONTRIBUTIONS 84 7.3 Con ibu ions made by Ha old Luis Pascua Cajucom I s a ed by spea heading he esea ch on he di e en ypes o dashboa ds a ailable, as well as he examples o each one. I ini ially ound ha mos dashboa ds a e used in businesses and ha he classi ica ions in ha con ex do no wo k wi h he con ex o moni o ing COVID-19 and his led me o analyze he exis ing dashboa ds based on hei common ac o s: geog aphic scope and ype o da a used. A e ha , I esea ched on he ypes o so wa e ools a ailable, especially o c ea ing and con igu ing dashboa ds. I also con ibu ed o he esea ch o da a o in e es , namely he exis en da abases o in e es and he g aphs o in e es . These we e da abases and g aphs ha may be pu in o conside a ion o inclusion in he dashboa d. In addi ion, I also esea ched on some a iables o in e es : R0, DI, and Inc. As o he da a p epa a ion s age o he p ojec , I ocused on e alua ing he da ase s, checking o poin s ha would become p oblema ic o conce n- ing when in eg a ed wi h he dashboa d o display da a. Fo he ini ial planning o he implemen a ion o he p ojec , I con ibu ed in deciding he di e en speci ica ions. Fo example, we decided o apply he Sc um me hodology, as we had some expe ience wi h his me hodology om ou expe ience in lea ning So wa e Enginee ing. We also decided a his poin o implemen he dashboa d using he RS udio IDE, oge he wi h Shiny and he o he RS udio lib a ies used. Finally, we decided o embed he dashboa d on o a Google Si es websi e, as we wan ed a websi e ha could easily be con igu ed so we could ocus mo e on he de elopmen o he dashboa d. Also in he planning s age, we decided on wha ype o COVID-19 dash- boa d we wan ed he inal p oduc o be. Based on he in o ma ion we ecei ed om he esea ch on he di e en ypes o dashboa ds, we decided o mainly ocus on epidemiological da a wi h a global scope, as we wan ed he dashboa d o be eachable o as many people as possible. Th oughou he de elopmen phase, I con ibu ed o all de elopmen sp in s by es ing and debugging he dashboa d and by o mula ing he con- clusions o each sp in . I was h ough his debug p ocess ha we we e able o de ec he anomalies ound du ing de elopmen ha would ha e o be esol ed by he ollowing sp in s. I also con ibu ed in de eloping he dashboa d by implemen ing he g aphs, CHAPTER 7. INDIVIDUAL CONTRIBUTIONS 85 especially he box plo s ha was allowed u he analysis o a iables. By he end o his phase, I managed o e ac o he sou ce code o imp o e i s comp ehensibili y. This includes de ining he auxilia y unc ions ound in he unc ions.Rsc ip , such as he c ea ion o ables and e ching o in o ma ion om he da ase . In he p esen a ion o esul s, I p epa ed he eposi o y ha con ained he da ase s used o he dashboa d, making hose da ase s a ailable o public use. Addi ionally, I c ea ed he use manual ha would help o ien a e he use s on he di e en a ailable unc ions o he dashboa d and acili a ed he example analysis o he si ua ions o Spain and he Philippines. This would help acili a e he use s in knowing how o use he dashboa d o ge he esul s hey need and p ope ly in e p e ing hose esul s o ge insigh s. Toge he wi h my colleagues, we o mula ed he conclusions made pos - de elopmen . We mainly ocused on he hings we lea ned om he di e en ields o dashboa d de elopmen , which includes p epa ing da a, adap ing o using di e en lib a ies and echnologies, and da a analysis. We also c ea ed he ecommenda ions o u u e wo k, some o which we e ac ions ha we ini ially wan ed o ake bu we e unable o due o ime cons ain s. Finally, I was in cha ge o asks ela ed o he documen a ion o he p ojec . This included he compila ion o sou ces ci ed in his wo k in o he bibliog aphy and p oo eading he wo k in gene al, making su e ha i was ee o any g amma ical e o s and ha e e y de ail abou he p ojec was well-explained. Chap e 8 Conclusions and Fu u e Wo k 8.1 Conclusions A dashboa d was c ea ed ha can in o m use s abou he p og ess o he pandemic in g aphic and map o ms, especially unde indica o s such as ansmission index, dange index, and cumula i e incidence. Th oughou he p ojec , he eam has used R and Shiny, as well as li- b a ies such as Plo ly and Lea le o de elop he dashboa d. Balsamiq Wi e- ames was also used o c ea e wi e ames and p o o ypes o he design o bo h he dashboa d and he websi e. Addi ionally, LaTeX was lea ned o he c ea ion o his wo k. Applying he Sc um me hodology in his p ojec was also an impo an p ac ice lea ned. As Sc um has become a s anda d in mos companies, lea n- ing how o employ his Agile me hodology in a p ac ical se ing p o es o be use ul in p epa ing o wo k unde employmen . Ano he c ucial skill aken om his p ojec is he applica ion o a ious analy ical echniques. Ha ing mul iple ypes o g aphs p esen in he dash- boa d allowed o he unde s anding o di e en aspec s o he da a (e.g. i s e olu ion o e ime, a iance in alues). Explo a o y da a analysis also plays a pa due o he na u e o he dashboa d o being based on a ma hema ical model. As he sou ces o da a ound in he da ase mainly consis o simple eco ded alues such as posi i e cases and dea hs, indica o s such as he R0, DI, and Inc we e ma hema ically de i ed om hese eco ds. The inal conclusion is ha , as impo an i is o analyze da a ho oughly and ex ensi ely, making hese da a unde s andable and accessible o he pub- 86 CHAPTER 8. CONCLUSIONS AND FUTURE WORK 87 lic is equally impo an . The use s o he dashboa d will consis o people o di e en backg ounds and o di e en pe spec i es on he pandemic. Hence, i is essen ial ha he same idea is objec i ely impa ed o hem by he dashboa d o a ain i s pu pose. 8.2 Fu u e Wo k Despi e ha ing success ully c ea ed a ully unc ioning COVID-19 dashboa d, he eam s ongly belie es ha he p ojec s ill con ains oom o imp o e- men . The i s ecommended ac ion is he au oma ion o he da a acquisi ion p ocess. Cu en ly, he dashboa d uses only one da ase , meaning ha he da a used could easily become obsole e in a ma e o days i he da ase is no cons an ly upda ed. Manually upda ing his da ase by uploading a la e e sion a e a gi en ime in e al is a iable ye s enuous op ion in esol ing his issue. Thus, au oma ing his p ocess will allow he dashboa d o s ay up o da e wi hou equi ing a la ge amoun o e o om he eam in e ms o main enance. Ano he ecommended ac ion is inco po a ing anima ions in o he map iew o he dashboa d, which could ampli y i s da a isualiza ion capabili ies. This would allow use s o see he e olu ion o he pandemic in di e en coun ies mo e clea ly, p o iding u he in e ac ion o he applica ion. Bibliog aphy [1] Wo ld Heal h O ganiza ion. Co ona i us disease (co id-19). h ps:// www.who.in /eme gencies/diseases/no el-co ona i us-2019. Ac- cessed: Oc . 30, 2020. [2] Wo ld Heal h O ganiza ion. O igin o sa s-co -2, 26 ma ch 2020. Tech- nical documen s, 2020. Accessed: Oc . 30, 2020. [3] Emilio de Beni o. Nea ly one- hi d o all in ensi e ca e beds in spain occupied by co ona i us pa ien s. h ps://english.elpais.com/ socie y/2020-11-04/nea ly-one- hi d-o -all-in ensi e-ca e- beds-in-spain-occupied-by-co ona i us-pa ien s.h ml, No 2020. Accessed: No . 26, 2020. [4] Ashley Killough and Oma Jimenez. ’bus ing ou o he seams’: Wes exas hospi als pushed o he limi in unp eceden ed co id- 19 su ge. h ps://edi ion.cnn.com/2020/11/19/us/wes - exas- co id-19-hospi aliza ions-su ge/index.h ml, No 2020. Accessed: No . 26, 2020. [5] Lo a Jones, Daniele Palumbo, and Da id B own. Co ona i us: How he pandemic has changed he wo ld economy. h ps://www.bbc.com/ news/business-51706225, Jun 2020. Accessed: No . 26, 2020. [6] Wo ld Heal h O ganiza ion e al. Co id-19 weekly epidemiological up- da e. 2020. Accessed: No . 26, 2020. [7] Sand a Du ce ic. Types o dashboa ds: S a egic, ope a ional & an- aly ical. h ps://www.da apine.com/blog/s a egic-ope a ional- analy ical- ac ical-dashboa ds/, Jul 2020. Accessed: No . 9, 2020. 88