Urban Public Transport Services in Selected Czech Towns

GeoScience Enginee ing Vol. 69 (2023), No. 1 geoscience.cz pp. 56–68, ISSN 1802-5420 DOI 10.35180/gse-2023-0089 URBAN PUBLIC TRANSPORT SERVICES IN SELECTED CZECH TOWNS Jiří HORÁK , Pa el KUKULIAČ , Ondřej KOLODZIEJ VSB – Technical Uni e si y o Os a a, Facul y o Mining and Geology, Depa men o Geoin o ma ics, Os a a, Czech Republic E-mail: [email p o ec ed] ABSTRACT U ban public anspo a ion sys ems in 26 selec ed Czech municipali ies a e s udied using hei ime ables, loca ion o s ops, and p ice su ey. The me hodology o hei p ocessing is p oposed. A se o indica o s desc ibing spa ial accessibili y, pe o mance, and p ices a e analysed mainly o hei ela ionships o he popula ion wi h an indica ion o he municipali ies which mos de ia e om common ends. A p oposed ou -dimensional e alua ion sys em co e s he spa ial accessibili y ope a ionalised by he numbe o s ops, he se ice equency using he numbe o ips pe ou e, he eloci y o selec ed anspo mode and he cos assessed by he p ice o a one-yea icke o equi alen . The esul s show he a e ees do no co espond o he pe o mance o u ban public anspo . Keywo ds: Clus e analyses; Czechia; Public anspo ; Reg ession analyses; Time schedule; U ban anspo . 1 INTRODUCTION U ban Public T anspo (UPT) ep esen s he mos impo an mode o public anspo acco ding o he numbe o passenge s. In 2015 UPT anspo ed 2160.8 mil passenge s in Czechia, which accoun ed o 80% o he o al public anspo (PT) and 44% o he o al passenge anspo [1]. The dominan ole o UPT wi hin PT pe sis s, and he di e ences in he las i e yea s a e no signi ican . The sha e o passenge anspo pe o mance (pskms) in UPT ep esen s almos 37% o all PT and 14% o he o al passenge pe o mance; he smalle pe cen age is due o sho e ips. Despi e he doub less impo ance o u ban public anspo sys ems (UPTS), he UPT s ill s ays away om he main in e es o geog aphe s. While na ional o egional public anspo e alua ions a e ela i ely equen opics, especially when discussing poo anspo se iceabili y o u al and pe iphe al a eas [2–4], he e alua ion and compa ison o UPTS a e missing. Time able da a a e commonly used o PT passenge ou ing. Seconda ily, hey can also be u ilised o analysing he s uc u e and o ganiza ion o PT ne wo ks [5–7], compa ing how PT ne wo ks a e s uc u ally o ganized ac oss ci ies [8–11] as well as he accessibili y and he le el o se ice hese ne wo ks p o ide [7, 12–14]. U iliza ion o public anspo ables o geog aphical esea ch pu poses is no a e, bu hey a e usually used o calcula e equencies o ansi connec ions o ehicles in selec ed ime pe iods be ween O igins and Des ina ions o measu e he in ensi y o in e ac ion. Michniak [15] u ilizes he equency o links o analyse in e na ional linkages o dis ic s in Slo akia as a p oxy o he in ensi y o in e ac ions. Ma ada e al. [16] e alua e public anspo ime schedules o he yea s 2000 and 2006 using he weigh ed equency o connec ions. The usage o u ban public ime ables is qui e a e. The spa ial accessibili y o u ban PT s ops in České Budějo ice was s udied by K a [17]. The accessibili y o supe ma ke s and hype ma ke s in B a isla a was e alua ed by 57 GeoScience Enginee ing Vol. 69 (2023), No. 1 geoscience.cz pp. 56–68, ISSN 1802-5420 DOI 10.35180/gse-2023-0089 K ižan [18], who uses a el ime be ween each u ban dis ic and supe ma ke s calcula ed om ime ables o e alua e accessibili y in he u ban en i onmen using eigh di e en ypes o accessibili y measu es. Ho ňák e al. [19] ha e s udied di ec ain and bus links om he ime ables o 1989, 1999 and 2011 o analyse he accessibili y and changes o accessibili y o eigh Slo ak u ban cen es. Nykl e al. [20] e alua ed u ban anspo accessibili y using ime ables o public anspo analysis, an inno a i e Dijsk a’s algo i hm, and comple e maps o oad, oo pa h and cycleway ne wo ks. The aim o his s udy is o demons a e he abili y o elabo a e a se o indica o s o an e ec i e quan i a i e e alua ion o UPTS om ime ables, o e alua e and compa e selec ed municipali ies o a ious sizes ac oss Czechia and o e eal o wha ex en he le el o UPT se ices de e mined by he municipal size. I a ouses he ollowing esea ch ques ions: How o e alua e he le el o UPT se ices? Is he popula ion he main in luencing ac o ? Does he p ice o cus ome s co espond o he le el o UPT se ices? 2 EVALUATION OF UPTS The e alua ion o UPTS a ies om he mos complex and mul idimensional cha ac e is ics o sho lis s o selec ed ac o s sui able o speci ic pu poses. Basically, we can dis inguish objec i e and subjec i e indica o s whe e he la e g oup e e s o a di ec e alua ion o a i udes, eelings and opinions o passenge s, which a e usually analysed om su eys, social ne wo ks o moni o ing o a el beha iou . Ob iously, subjec i e ac o s e lec some le el o he objec i e c i e ia ha may be unde s ood as p ima y ac o s. The complexi y o e alua ion is ela ed o ou e o o c ea e a sus ainable u ban anspo sys em (SUTS). I equi es s eng hening a ious sys em ea u es, including accessibili y and mobili y, eliabili y and e iciency, as well as sa e y and secu i y, social equi y, con enience and com o [21]. Add essing all ea u es and c ea ing a uni e sal e alua ion is a e y ha d ask. Fo example, Ľup ák e al. [22] speci ied he quali y o he p o ided anspo se ices by a se o he ollowing aspec s: egula i y, eliabili y, sa e y, speed, economy, easonable cos , ecology, com o , pe o mance and a ailabili y; ne e heless, he p oposed ac o s we e measu ed by desc ip i e (quan i a i e) indica o s like se e al ypes o ime, dis ances, speeds and he numbe o connec ions, whe eas he only quali a i e aspec was he ype o ain. Al Mamun and Lownes [23] p o ided a mo e conc e e se o ecommended measu es: Se ice Co e age, Time-o -Day, Wai ing Time, Se ice F equency, Demog aphic da a, Vehicle Capaci y, Rou e Co e age, T a el Time, T a el Cos , Hou s o Se ice, Walking Rou e, Access dis ance, Com o & pa king, and Ne wo k connec i i y. O he au ho s a e ocused only on selec ed c i e ia which, acco ding o hei opinions, dominan ly in luence he e iciency and u ilisa ion om he use ’s poin o iew. Fo example, acco ding o Cede e al. [24], an e ec i e public anspo se ice can be de ined as minimum in- ehicle a el ime and wai ing ime; A ila-To es e al. [25] emphasize also low a e in addi ion o less a el ime and less wai ing ime. Some au ho s ad oca e he key ole o di ec connec ions in public anspo u ilisa ion, e.g., K izan e al. [26] in he complex e alua ion o he accessibili y o supe ma ke s in B a isla a o Ho ňák e al. [19] o he e alua ion o in e egional communica ion. Ano he app oach is o apply measu es based on he g aph heo y. B inke [27] ecommended he use o ne wo k de iali y, ne wo k densi y ( ela ed o bo h he a ea and he popula ion), and connec i i y. We can summa ize ha he ollowing key objec i e ac o s a e conside ed: 1. Spa ial accessibili y o he anspo – app op ia e me ics a e he ne wo k densi y, numbe o s ops, s op densi y, ope a ional leng h o a e age dis ance o s ops. Fo he in eg a ion o une en spa ial dis ibu ion o sou ces and des ina ions, i is ecommended o calcula e also sha es o popula ion, buildings, o POIs well accessible om s ops. 2. Se ice equency – ela ed o wai ing ime, wi h app op ia e indica o s o he numbe o depa u es pe s op o he numbe o ips pe ou e. Al e na i ely, he numbe o ips pe day can be applied as a joined indica o o bo h accessibili y and equency. 3. Veloci y – ela ed o in- ehicle ime, exp essed di ec ly by he a e age speed. 4. Fa e – exp essed by he annual subsc ip ion (p ice pe yea ). We can conclude he ideal UPTS has high accessibili y (high ne densi y, low dis ances), high se ice equency (low wai ing ime, high ips pe ou e, high ips, high depa u es), high speed, low a e, high a iabili y and low 58 GeoScience Enginee ing Vol. 69 (2023), No. 1 geoscience.cz pp. 56–68, ISSN 1802-5420 DOI 10.35180/gse-2023-0089 de iali y. Co esponding s uc u al and pe o mance indica o s should be a base o a comp ehensi e quan i a i e e alua ion o UPT. A c i ical aspec is ime. A a ia ion o indica o s du ing a day o week cycle should be included in he e alua ion. In his s udy, empo al changes a e e lec ed by indica o s enume a ing di e ences be ween se ices on Wednesday (Wed.) and Sunday (Sun.). 3 DATA SOURCES AND METHODOLOGY In o ma ion abou he geog aphical ex en o UPTS in Czechia is no easily a ailable. We can e alua e he s a us acco ding o he lis o downloadable u ban ime ables on he CHAPS ( he adminis a o o he Na ional In o ma ion Sys em on Regula Public Passenge T anspo Time ables) websi e. O 6253 municipali ies (middle 2015), 103 decla e o ha e hei own u ban anspo . O cou se, many o he municipali ies, especially on he ou ski s o la ge ci ies, a e se ed by a public anspo company ope a ing om he cen al ci y. Fo example, P ague Public T anspo Company Inc. ope a es in 39 o he municipali ies, B no Public T anspo in 23, and Os a a T anspo in 20 (No embe 25, 2015). The expansion o he su oundings con inuously g ows, especially in he case o P ague, whe e, since 2015, new ex e nal zones 8 and 9 we e added. The e alua ion is complica ed by joining UPTS in o egional PT sys ems, which may co e he ull NUTS2 o an e en la ge a ea (Fig. 1). The in eg a ion o PT se ices in he case o B no UPTS co e s he ull Sou h-Mo a ian egion. Figu e 1. Ex ension o in eg a ed UPTS ope a ed o 26 selec ed municipali ies in Czechia. Sou ce: au ho s’ elabo a ion The numbe , ex en and mix u e o UPTS complica e compa a i e analysis. App op ia e da a is no publicly a ailable and no ee o cha ge. Al hough s ill mo e and mo e wo ldwide ci ies p o ide hei ime ables in GTFS o ma and new s anda dised da a se s a e a ailable [28], he local PT si ua ion in smalle owns is hidden and di icul o quan i y, measu e and e alua e. Howe e , in such places, a ious issues can be an icipa ed. The high numbe o passenge s, ich subsidies and long- e m con inual p og ess in la ge ci ies guide owa ds well-de eloped UPTS. On he con a y, smalle municipali ies mee di icul ies in how o assu e a good le el o public anspo se ices a a easonable cos . 59 GeoScience Enginee ing Vol. 69 (2023), No. 1 geoscience.cz pp. 56–68, ISSN 1802-5420 DOI 10.35180/gse-2023-0089 The limi ed inancial sou ces and in e es o go in o de ail in some aspec s equi ed o limi he numbe o analysed municipali ies o 26, which ep esen s app oxima ely one qua e o all municipali ies o ganising hei own UPTS in Czechia. The selec ion co e s all egional capi als and 13 smalle owns. Among a ious possible selec ion c i e ia, local elie is conside ed one o he impo an ac o s whe e la ge e ical change should complica e he se ice, he ne cons uc ion and he speed. Hence, he digi al model o elie om A cCR500 o e laid municipal polygons, he minimal and maximal ele a ions wi hin polygons we e eco ded and he ollowing owns we e selec ed: municipali ies wi h a high ele a ion and high a iabili y (V chlabí, Děčín, Os o , Vimpe k, T u no and Jablonec nad Nisou), low a iabili y and/o low ele a ion (Ha ířo , Dačice, Teplice, Kladno, S říb o, Uhe ské H adiš ě and Bys řice nad Pe nš ejnem). Da a on u ban anspo ime ables om 26 selec ed Czech owns was bough in 2016. Full- ime schedules, incl. he loca ion o e e y pos o u ban anspo s ops, we e expo ed o 2 days, No embe 25 (Wednesday) and No embe 29 (Sunday), 2015. Da a does no con ain ime ables o any anspo , bu only hose assigned o u ban anspo sys ems. The impo an ques ion was how o es ablish an app op ia e ange o analysis. A UPTS can con ain se ices o di e en p o ide s (including Czech Railways, e.g.) and usually co e s no only he cen al own bu also su ounding municipali ies, e lec ing he equi ed su icien anspo connec i i y wi h he mic o egion co e. I is ob ious ha only la ge municipali ies es ablish hei own public anspo company o p o ide UPTS. Such se ices in smalle owns a e usually gua an eed by companies p o iding se ices in he egion (i.e. KAD o V chlabí, Osnado o T u no ), in he g oup o ci ies (usually pai s like Libe ec-Jablonec n.N., Zlín-O oko ice) o , on he con a y, by small p i a e companies (Zla o ánek o Bys řice n.P. o Jose Š e l Tou o Dačice). The s udy is ocused on all PT se ices wi hin he adminis a i e bounda ies o he main municipali y, including all se ices ( om di e en p o ide s) o e ed o cus ome s in he ame o he uni ied u ban a e ees (Fig. 2). The basic idea is o e alua e how many se ices a e p o ided o cus ome s, in o he wo ds, how e icien he sys em is om he use ’s poin o iew. Only some o he ele an ques ions we e e alua ed o he se ice o he main p o ide (i.e., a e age eloci y). Figu e 2. UPT ne wo k in 26 selec ed municipali ies in Czechia. Sou ce: au ho s’ elabo a ion The municipal popula ion was calcula ed by agg ega ing he popula ion om he las census (2011) assigned o buildings (da a RÚIAN, 2016) wi hin app op ia e municipal polygons. In he case o a sepa a e municipal a ea, 60 GeoScience Enginee ing Vol. 69 (2023), No. 1 geoscience.cz pp. 56–68, ISSN 1802-5420 DOI 10.35180/gse-2023-0089 only he main a ea was selec ed. The popula ion densi y was calcula ed as he a io o his popula ion and he ele an a ea. Clea ly, he delimi a ion o he ci y by adminis a i e bo de s is no sui able o all pu poses. A di e en app oach was applied by Kujala e al. [28], who de ined o each ci y a cen e-poin and selec ed he ci y a ea by spa ial il e ing wi h he speci ied bu e adius ( om 10 o 50 km, acco ding o he ci y). Da abase p ocessing played a key ole in da a managemen , con ol and analysis. The o iginal da a (CSV o ma , 126 MB) was impo ed in o he da abase. The p e-p ocessing was ocused on e o de ec ion and co ec ion o e o s o assu e da a consis ency du ing calcula ions, selec ion o ep esen a i e ips, da a agg ega ion, and in e sec ion wi h a egis y o buildings. The ex en o p ocessing may be documen ed by he di e ence be ween he o iginal da a and he inal size o he da abase, which is 2161 MB. G ea a en ion has o be paid o e o de ec ion and epai . The da a is no e o -p one and con ains a ious kinds o issues (e o s, imp ecisions – imp ecise eco ding o me age when in o ma ion sys ems o some p o ide s eco d only kilome es and no me e s, inaccu a e loca ion o s ops, and duplica es – epea ed ou es and ips due o o e lapping expo s o di e en ci ies o di e en schedule es ic ions). Especially duplica es ha e o be ca e ully elimina ed; o he wise, esul ing s a is ics will be signi ican ly biased. The se iousness o his p oblem can be seen om he e idence ha in he o iginal da a se almos 2/3 o eco ds a e duplica ed. The mul ile el con ol sys em has o c oss-check any o m o inconsis encies in da a o de ec po en ial p oblems. I was necessa y o b eak down e e y ou e o edges ep esen ing elemen a y mo emen be ween wo consecu i e s ops. Fo each edge, depa u e imes, a i al imes, and cumula i e me age we e eco ded. F om his p ima y da a, he a el ime, a el and Euclidean dis ances we e calcula ed. The speci ic issue is how o deal wi h imp ecise eco ds whe e he cumula i e dis ance (me age) is eco ded, wi h p ecision limi ed o km (no decimal places). Wi hou co ec ions, he da a consis s o edges o ze o leng h, which causes e o s in local eloci y calcula ions. Fu he con ol checks i he a el dis ance is la ge han he Euclidean dis ance. To sol e hese issues, a 3-s ep da a adjus men sys em was adop ed. Fi s , all s ops wi hin 1 km (all eco ds wi h he same me age = cumula i e dis ance) we e equally dis ibu ed wi hin his kilome e. Second, he mo ing a e age was applied o smoo h sudden changes be ween edges. And inally, he dis ances we e adjus ed o be equal o o la ge han he Euclidean dis ance. In such a way, da a consis ency is assu ed. Ano he ype o con ol calcula es local eloci ies and checks (and epai s) all un ealis ic alues (speed). The p e-p ocessing also includes p epa a o y da a modi ica ions. Fo some asks, a ep esen a i e ip o each link ( ou e) was selec ed (examples in Fig. 3). I was assumed he longes ip in wo king days would bes ep esen he ou e. Using he ep esen a i e ip, he ope a ional leng h was enume a ed. The second main p epa a o y ask was o c ea e anspo s ops by agg ega ion acco ding o he name o he s op om he o iginal se o all anspo pos s. Figu e 3. Simpli ied anspo ne wo k in H adec K álo é (le ) and České Budějo ice ( igh ). Sou ce: au ho s’ elabo a ion. S ops and ou es modi ied om CHAPS L d. da a 61 GeoScience Enginee ing Vol. 69 (2023), No. 1 geoscience.cz pp. 56–68, ISSN 1802-5420 DOI 10.35180/gse-2023-0089 3.1 Spa ial accessibili y o PT The alues o he main indica o s o e alua ing spa ial accessibili y o all explo ed municipali ies we e enume a ed. We can no e ha he numbe o s ops and he ope a ional leng h di ec ly depend on he municipal size, while o he s a e conside ed ela i e o neu al indica o s. The numbe o s ops indica es he accessibili y o UPTS and i s se ices. I is belie ed ha a highe numbe (o densi y) o s ops is mo e ad an ageous, especially in he case o local ips. Ob iously, he numbe o s ops depends on he a ea and popula ion o he municipali y. The eg ession model sugges s a e y high co ela ion (R2 o he quad a ic equa ion is 0.994) be ween popula ion and he numbe o s ops (Fig. 4). Using 5% con idence in e als, we can dis inguish an unusually high and low numbe o anspo s ops. An exceeding numbe o s ops on wo king days can be seen in Libe ec, Ús í nad Labem, Olomouc, Zlín, Ka lo y Va y, Děčín and Jablonec nad Nisou. On he opposi e, an abno mally low numbe o anspo s ops a e a ailable o esiden s in České Budějo ice, Ha ířo , Kladno and Uhe ské H adiš ě. Figu e 4. Numbe o s ops ela ed o popula ion (le ) and popula ion densi y ( igh ). Sou ce: au ho s The desc ip ion o he si ua ion can be well supplemen ed by mapping he ela ionship be ween densi ies. Jablonec has he highes densi y o s ops o medium popula ion densi y. I s ex eme posi ion is clea ly po ayed in he g aph o he ela ionship be ween he s op densi y and popula ion densi y (de ail in ig. 4). The ela i ely high s op densi y is in Ka lo y Va y, while e y low densi y is in P ague, Ha ířo and Uhe ské H adiš ě. The empo al di e ences show hese indica o s a e s ongly ime-dependen . Especially in small owns, we can see a d ama ic educ ion in he numbe o s ops be ween wo king and non-wo king days. On wo king days, UPTS p o ides se ices o p ac ically all s ops, bu on weekends he numbe is educed. Di e ences in he numbe o s ops a e negligible o la ge owns (usually se ice declines by 2–9%), bu small owns signi ican ly educe hei se ice in ex eme si ua ions o ze o (Vimpe k, Dačice). Mo e han a 10% educ ion is me in S říb o, Jablonec nad Nisou, T u no , Os o and Bys řice nad Pe nš ejnem. The ex en o he UPTS can be ela ed o he ope a ional leng h (Fig. 5). The ope a ional leng h ep esen s he o al leng h o all ou es inside a ci y o all p o ide s wi hou duplica e edges. I should be aken in o accoun ha he esul is di e en om he ope a ional leng h p o ided by he main UPTS p o ide s. This analysis includes only pa o UPT loca ed inside he ci y and o all p o ide s. Abno mally high (abo e 5% con iden ial in e al) ope a ional leng h inside a municipali y is in Libe ec, in con as o he unusual low in České Budějo ice. Qui e low ope a ional leng h (bu no signi ican ly low) is in Ha ířo , Teplice, and Kladno. 62 GeoScience Enginee ing Vol. 69 (2023), No. 1 geoscience.cz pp. 56–68, ISSN 1802-5420 DOI 10.35180/gse-2023-0089 The ne densi y (Fig. 5) was calcula ed as he sum o unique anspo edges inside he ci y di ided by he o al a ea o he ci y. Simila esul s o ope a ional leng h can be ound o he eg ession analysis wi h he ne densi y. Excep ionally high ne densi y is in Jablonec nad Nisou; a high alue is also ound in Libe ec, while Teplice o Kladno ha e e y low ne densi ies. Figu e 5. Ope a ional leng h ela ed o popula ion (le ) and ne wo k densi y ela ed o popula ion densi y ( igh ). Sou ce: au ho s The a e age dis ance be ween s ops ep esen s a equen ly used indica o ela ed o he accessibili y o PT sys ems. Sho e dis ances imp o e he accessibili y o he se ice; ne e heless, i nega i ely in luences he a e age speed o he anspo . Sho dis ances be ween s ops (Fig. 6) can be ound in No he n Bohemia (Ús í n.L., Teplice, Jablonec n.N.) (abou 540–550 m). Dis ances abo e a e age a e in smalle owns: V chlabí, Vimpe k, Uhe ské H adiš ě, and S říb o (a ound 900 m). Figu e 6. A e age dis ance be ween s ops ela ed o popula ion densi y. Sou ce: au ho s 63 GeoScience Enginee ing Vol. 69 (2023), No. 1 geoscience.cz pp. 56–68, ISSN 1802-5420 DOI 10.35180/gse-2023-0089 3.2 Se ice equency The basic e alua ion o he anspo equency can be eached om he numbe o ips pe day, bu i should be conside ed i is a join indica o co e ing pa ly bo h equency and accessibili y. A eg ession analysis (Fig. 7) be ween he popula ion and he numbe o ips pe day p o ides mo e simple esul s han in he case o he numbe o s ops. All owns excep B no show an expec ed numbe o ips. B no has a highe numbe o ips pe day han expec ed. Mo e han a 50% empo al dec ease can be seen in all explo ed owns wi h popula ions below 50,000 excep o Ka lo y Va y (only a 29% dec ease). F om la ge owns, České Budějo ice belongs o his g oup (–52%). F om mid-size owns, we can also no ice Olomouc has a e y small dec ease in se ice on Sunday (only a 30% dec ease). Simila esul s can be ob ained om he e alua ion o all ips (no only inside he ci y) whe e he pe cen age dec ease is he same, excep o h ee owns, B no (–55%), Jablonec (–61%) and Bys řice (–74%). Figu e 7. Numbe o ips pe day (le ) and pe ou e ( igh ) ela ed o popula ion. Sou ce: au ho s The main equency indica o , he a e age numbe o ips o one ou e, shows how ou es a e ac i e. The high a e age numbe o ips pe ou e is ypical o la ge ci ies; wi hin hem, Plzen is he leade , while P ague occupies he second place. This big e ad is ollowed by some mid-size egion capi als like České Budějo ice ( he same numbe as Os a a), Jihla a, Teplice, Olomouc and Ús í nad Labem. On he opposi e, Dačice and Vimpe k ha e he weakes ac i i y on ou es. F om he ela ionship be ween popula ion densi y and an a e age numbe o ips pe ou e (Fig. 7), we can see weak ac i i y o Ha ířo , Kladno, Jablonec nad Nisou and Uhe ské H adiš ě, and a good equency ela ed o popula ion densi y in Plzeň and Jihla a. I we compa e he a e age numbe o ips pe ou e on Wednesday and Sunday, we ound qui e a wide ange o mani es a ions om 0-change (Jablonec nad Nisou), e y small (less han 20%) in Olomouc, Libe ec, H adec K álo é; h ough usual decline a ound 20–30% o a se ious d op down mo e han 50% o Bys řice nad Pe nš ejnem, S říb o, T u no , Uhe ské H adiš ě, V chlabí, and wo known owns wi h no ope a ions on Sunday (Dačice, Vimpe k). Ano he equency indica o , he a e age numbe o depa u es om each s op pe day, shows e en be e he posi i e ela ionships o he popula ion size (Fig. 8). The ela ionship is almos linea o smalle owns (less han 64 GeoScience Enginee ing Vol. 69 (2023), No. 1 geoscience.cz pp. 56–68, ISSN 1802-5420 DOI 10.35180/gse-2023-0089 150,000 inhabi an s). Fu he inc ease in he a e age numbe o depa u es wi h popula ion is mode a e and con i ms he o de o popula ion size o he ou la ges ci ies. Figu e 8. A e age numbe o depa u es om each s op pe day ela ed o popula ion (le ) and popula ion densi y ( igh ). Sou ce: au ho s The ela ionship o popula ion densi y (Fig. 8 igh ) is weake (R2 only 0.542), and de ia ions o owns om he linea ela ionship a e no signi ican , e en hough we can see some di e ences among owns. I.e., Ha ířo , Kladno, Jablonec, Uhe ské H adiš ě demons a e a lowe numbe o depa u es han wha is expec ed by he eg ession model. Tempo al changes (Wed.-Sun.) in he numbe o depa u es om each s op pe day a e no as a iable as he p e ious indica o . The lowes changes a e in Bys řice (bu due o he small size o UPTS, i is no a con iden esul ) and in Olomouc. The bigges changes (sha p dec ease o he se ice on Sun.) a e in Uhe ské H adiš ě, S říb o and Os o . 3.3 Veloci y Veloci y is an impo an aspec o anspo connec ion quali y. I is ecommended o e alua e he a e age eloci y in a municipali y acco ding o he anspo mode. The compa ison is based on bus eloci ies, which is he only PT mode a ailable in all municipali ies. When compa ing u ban bus eloci y, he lowes speed is in Ka lo y Va y and Jablonec n. N. (abou 20 km/h), which co esponds wi h he idea o e ically a iable and na ow s ee s and he occu ence o geog aphical ba ie s ( i e s, e c.). Ne e heless, he ela ionship o he e ical changes is no simple (Fig. 9). The highes ele a ion di e ences can be seen o Děčín wi h below-a e age bu no lowes eloci y. The la es elie is in Pa dubice and H adec K álo é wi h a nea -a e age eloci y. We may conclude ha local elie is less impo an han o he anspo ac o s. High eloci ies o u ban buses a e in Os o , Uhe ské H adiš ě, S říb o (smalle owns) and Plzeň and Os a a om he g oup o la ge ci ies.