Analysis Diagrams
Analysis diagrams are available in different formats:
- Histograms
- Corridor statistics
- Station statistics
- Delay evolution
- Comparisons
General structure of analysis diagrams
The analysis diagrams are structured in four main elements:
- the diagram itself
- the diagram type selection
- the data source settings
- the data filtering and visualisation options
The elements 2 to 4 are located in the sidebar.
Depending on the diagram type, different filters are available. Generally, the filters include:
- the corridor
- the direction
- the OCP
- the train categories
- the train codes
Some of the filters are obligatory, some are optional. Different statistics have different obligatory filters, but it will be highlighted if necessary values are missing. More details on train filters can be found in the section "Train filters" below.
The visualisation usually accepts user-defined thresholds in seconds or minutes and extreme values can be cut off.
Data source settings
In the analysis module of trenoLive it is possible to visualise either real data or simulation results. The real data integration needs to be organised by TRENOlab. Instead the integration of simulation results is automatic.
Simulation data as source data
The simulation results are directly accessed by trenoLive after the simulation finished in trenissimo. It is possible to analyse multiple different simulated scenarios based on the same timetable. The simulation data can be identified by the Code of the Simulation setup. Only simulation setups from the current scenario can be selected.
If a simulation has been restricted with a simulation area covering only a part of the macro model network, the corridor filter needs to be selected with attention, as some corridors might not be able to visualise any data.
For simulated data three different filter types are available:
- Available runs (limits the number of stochastic iterations taken into the evaluation)
- Start date / End date (limits the evaluation to certain days, only relevant for multi-day simulation)
- Days of the week (limits the evaluation to certain weekdays, only relevant for multi-day simulation)
The simulated data matches with the train codes of the planned timetable.
Real data as source data
The real data is imported via an API. Only one dataset can be analysed. A restriction of the analysis period via the Start date / End date filter or the Days of the week filter is possible.
Train filters
The diagram shows real or simulated data for all trains and entries matching the other citeria (corridor, direction, OCP, times etc.), if no train filters are imposed. The number in brackets (e.g. "Train [815]") shows how many elements can be filtered further, because they satisfy all the other filters. Train filters allow the user to further limit the results. Three filter modes are available at the moment:
- Train searches (filter for all trains matching with the train codes extracted by the train search, multiple train searches can be selected)
- Collections (filter for all trains matching with the train codes in the selected user or scenario collections, multiple collections can be selected)
- All trains (no further filter is applied)
Nonetheless, in all three modes the list of trains presented can be further limited manually selecting only specific: 4. Train categories 5. Tags 6. Train codes 7. Train numbers
Filtering by train numbers will extract all trains with the first train number matching the inserted filters. If the train number changes in a path segment, it will not be considered in this filter. We advise to use this filter only in cases where trains never change their train number during their itinerary. For cases where train number changes occur it is better to filter by train code, to aviod incomplete or wrong analysis results.
The diagrams in detail and how to use them
Delay Histograms
Delay histograms show the number of entries in an OCP or a list of OCPs that are in a certain delay intervall (e.g. between 60 and 120 seconds). The intervals can be user defined in the visualisation settings.
The delay histograms can also be used to show the number of absolute dwell times in an OCP or absolute run times between two OCPs that are in a certain intervall (e.g. between 60 and 120 seconds), without calculating delays against a scheduled timetable.
Track occupation
The analysis track occupation diagram shows both the planned and the actual occupation of platforms / tracks in an OCP with OCP tracks defined. The track occupation diagram can be understood as an equivalent to the Analysis timetable graph. The planned time is indicated by an empty rectangle with a coloured border, the actual times are shown as slightly transparent coloured histograms.
At the moment the diagram shows all planned trains, including overlapping variants, which can cause issues in the readability, if they have not been filtered. We are working to improve the transparency and will introduce a mechanism to agglomerate data for variants with the same / similar timetable. At the moment the Scheduled day is out of function.
Track bar timeline
The track bar timeline details the planned and actual platform / track occupation time for a single train over all of its operating days. It shows the planned time in a light red shading and and the actual times in black and dark grey bars (different shading for better distinction of individual values).
Track scatter plot
The track scatter plot shows the relationship between arrival and departure delay with dots for each individual entry (e.g. each day): Trains arriving and departing with the same amount of delay are on the line x=y. If the delay increased or decreased at the station the dots are above or below the line.
Track histograms
The track histogram is a subcategory of the delay histogram and displays the actual data in a similar format. It shows all arrival delays for a single train code in a single station.
Track pie chart
The track pie chart shows the share of entries on a specific OCP track. It can be used to determine if trains get replatformed frequently.
Station - Mean delay by train
The mean delay by train shows the mean delay over all entries for all trains arriving or departing in an OCP. It should be used to identify critical trains with high mean delays. Be careful about the number of data points, as trains with high delays and many operating days might be more critical than trains with high delays, but only a couple of operating days. The number of data entries is available in a tooltip on mouseover at each bar.
Station - Mean delay by time
The mean delay by time shows the mean delay over all entries for all trains as they are scheduled during the day. The x-axis is time during the day. Considerations are similar to the Mean delay by train diagram above.
Station - Percentile delay
The percentile delay diagram shows delay percentiles over all entries for all trains arriving or departing in an OCP. It also shows minimum and maximum delays for each train. The Percentiles can be user defined.
Station - Threshold delay
The threshold delay diagram shows delay percentages at user defined thresholds over all entries for all trains arriving, departing or stopping in an OCP.
Corridor delay evolution
The corridor delay evolution shows the evolution of train delays over a corridor. Each operating day represents a grey curve, the mean and medium delay as well as a Standard deviation intervall are indicated in colours. It can be used to identify line sections or stations where delay increases or decreases systematically. It can also be used to identify outliers, journeys with exceptionally high delays or exceptionally high delay increases in a certain section.
The diagram allows the visualisation of multiple train codes, but is of limited use for big data analysis. To utilise the delay evolution diagram attention must be paid to not select too few or too many data points.
Corridor percentile delay
The corridor percentile delay shows the delay percentiles over all entries for all trains arriving, departing or stopping in all (or selected) OCPs of a corridor. It allows to identify locations
Corridor threshold delay
The corridor threshold delay shows the delay percentages over all entries for all trains arriving, departing or stopping in all (or selected) OCPs of a corridor. It allows to identify locations on the corridor where the punctuality decreases or improves.
Delay report
Work in progress
Comparison of single trains
Work in progress
Comparison of delays by day
Work in progress