Increase lwd of bus lanes on map (width) to 3 or 5

Submit both to GET: * GTFShift * Master dissertation

Focus on EDA messages only and messages + events results

Add bus bunching to case-study

Disclaim that only 10

Overtake

Use grid instead of heatmap!

Add indicator per trip

Shortenings can complement bus bunching analysis and should be related with them, so no need for week/hour evolution.

Actually, there should be a chart with % shortenings vs % bus bunching, to study correlation between the two variables.

ALTERNATIVE: Difference between days is residual. If y starts on 0 and line is smoother, there won't be a visual difference throughout the week.

Since there is no statistical evidence of a tendency, the chart would be better explored in 4 quadrants (P50 dividing them), and exploring how the 4 groups are characterized

For each quadrant, choose some routes to explore in the text as examples, and for those provide chart with nr of buses bunched (y, comulative) per stop number (x)

ALTERNATIVE: try to normalize per route extension (m). Y would be % of services bunched per km and then there should be a tendency

This is a good value, since Carris average is around 7. Having a value close to the average means that the edge cases happen outside the bus lanes.

Also, compute for other big axis that do not have bus lanes, to validate that bus bunching occurs more there: * Morais Soares * Avenida do Brasil

It's ok to show only worst cases, but specify (top 5 or top 10)

Complement with histogram of % per stop (to affirm that for 25% of stops, 1 in every 10 buses is bunched)

Also disclaim % of stops (without services dimension)

Distinguish weekends of week days

Make line smoother

Start y axis on 0.0, and units to % (instead of decimal values)

Again: weekends vs week days

Add columns with route extension and commercial speed (extension/duration), for easier interpretation

Show also routes with low values

Add histogram with all routes, for people to have the big picture

CLARIFY: analysed, corresponding to

Add columns with route extension and commercial speed (extension/duration), for easier interpretation

CLARIFY: analysed, corresponding to

Data driven mngmt

• limitações tecnológicas prejudicam operação

Retirar ónus da decisão do motorista

Tipificar e automatizar respostas a disrupcoes no serviço como redundância á falta de resposta da central

And unexpected

That in some circunstances can be Critical (add examples)

Teste

Teste :)