Making Cycling Data Visible

Five primary datasets are sourced from Ville de Montréal and Quebec open-data portals:

Automated jobs keep the dashboard in sync with open data sources. Cycling data and cycling network layers are synchronized once a week. Cycling sync is incremental: after a one-time backfill (only if daily coverage since 2009 is incomplete), it refreshes only recent yearly archives and updates rows only when values changed. Construction data is refreshed daily — permits and street-level impacts are fetched, missing locations are resolved, and a daily snapshot is stored with a rolling retention window. SAAQ collision data is synchronized from yearly files and each year is refreshed transactionally so reruns stay deterministic.

The overview combines all daily cycling records within your selected date range. Total rides is the sum of all passages recorded across every counter. Active counters shows how many counters reported data during that period. The daily average is the total volume divided by the number of days in the range. Each KPI shows a trend badge comparing the current value against the same-length prior period (e.g., if you select 90 days, the comparison covers the 90 days before that). A "Data through" label indicates the latest date with available data. The daily chart also includes an optional weather-context overlay (precipitation, snowfall, and mean temperature) sourced from city-level daily weather observations.

Monthly volumes are grouped by calendar month and organized by cycling season (June to February). The chart overlays the current season against the previous one so you can spot differences at a glance. The values shown are the actual recorded totals — no averaging or smoothing is applied.

Each cycling counter is placed at its geographic location. Larger circles represent higher traffic volume within the selected date range, scaled so that both busy and quiet counters remain visible. A seasonal cycling-network overlay can be toggled between 4-season, summer, and winter layers. BIXI stations are overlaid from the live GBFS feed with current bike and dock availability. Construction markers always reflect the most recent daily data, regardless of the date filter. Sites where the bike path is fully blocked appear in red; partially obstructed sites appear in amber.

Cycling volume is grouped by borough based on each counter's location. Totals are ranked so you can see which boroughs have the most cycling activity. The chart shows the relative share of each borough within the selected date range. Note: volume reflects the number of counters installed in each borough — boroughs with more counters will naturally show higher totals. The tooltip shows the counter count for each borough to help interpret the comparison.

A grid showing month and day of the week, where each cell represents the average daily passages for that combination across all counters for the selected date range. The legend shows the actual value range (minimum to maximum). This helps reveal patterns like higher ridership on weekday commutes versus weekends, and summer peaks versus winter lows.

Only construction permits with a bike path impact other than "Aucun impact / non applicable" (no impact) are displayed. Each row is enriched with:

The safety page builds an accident co-involvement matrix and a monthly evolution chart from SAAQ annual collision reports. Rows represent the road-user profile involved in a collision; columns represent counterpart context. Because SAAQ open data is accident-level (not victim-level), one collision can contribute to multiple row-column cells when several user types are present. Filters support year range, severity, Montreal-only scope, and highway exclusion. Severe outcomes are represented by the source combined category ("Mortel ou grave"), and trend interpretation across 2013-2014 should be done with caution due to a structural break in reported material-damage records. This section is intended for prioritization and monitoring trends, not legal fault attribution.

You select a compared month and a baseline month. For each counter, monthly totals are shown for both months, while percent change is calculated from average daily volume (monthly total divided by calendar days) so month-length differences are normalized. Rows can be explored through ranking views (largest increases, largest decreases, most stable), filtered by borough/construction/thresholds, and expanded to show a 12-month trend with highlighted compared and baseline points. Days-with-data coverage is displayed to flag lower-confidence comparisons.

When a counter is selected, its daily volume is shown for the last 30 days (or the selected date range). You can also toggle weather context on the chart (temperature, precipitation, snowfall, and wind) to support day-level interpretation. Any active construction within roughly 1 km is listed along with its estimated impact on cycling.

To keep the dashboard fast, data is refreshed periodically rather than on every visit — cycling data weekly, cycling network layers weekly, construction data every 15 minutes, and BIXI station status every minute. Since the construction source is updated once per day and cycling/network sources are updated weekly, more frequent refreshes would not produce newer results.