HSE Incident Data Analysis

This page has the full build guide — the data you need, the Power Query steps, the model, the measures (the DAX, written out), the visuals, and the schedule — plus a live demo chart below with a sample 12-month dataset: switch between the monthly trend, the incidents by type, and the near-miss ratio, and download the CSV to start your own. The KPI formulas are the section to bookmark.

How to use the tool

1. Gather the data — the incident register (the date, the type, the severity, the lost time, the cost, the department, the near-miss flag), one row per event, the CSV or the sheet
2. Get it into Power BI — the Power Query editor: the types, the date column, the removed duplicates, the normalised text
3. Build the model — the incident table, the date table (the calendar), the relationship on the date; the star, not the snowflake, for the start
4. Write the measures — the TRIR, the LTIR, the severity rate, the near-miss ratio, the cost, the YTD: the DAX is written out below
5. Build the visuals — the trend line (the TRIR by month), the bar (the incidents by type), the KPI cards (the current, the target, the delta), the treemap (the department) — then schedule the refresh and publish

Worked example

The sample dataset is 12 months of a 240-worker plant: 7 recordables (the TRIR 2.9 against the 2.8 target — the one-month slip the committee is watching), 1 lost time (the LTIR 0.4), 96 near misses (the ratio 13.7:1 — the leading indicator is healthy), the top type the manual handling (3 of the 7), the cost line 48k. The dashboard: the TRIR line trending down through the year with the target line, the type bar with the manual handling leading, the KPI cards with the red/amber/green against the targets, the department treemap. The schedule: the refresh nightly from the SharePoint CSV, the row-level security per site, the email to the HS committee on the Monday. That is the incident report in Power BI — the data, the model, the measures, the visuals, the schedule.

Standard & legal context

The KPIs are the industry standard: the TRIR (the Total Recordable Incident Rate, per 100,000 hours — the OSHA 300 rate), the LTIR (the Lost Time Injury Rate), the DART (the Days Away, Restricted or Transferred, the OSHA pair), and the leading set — the near-miss ratio, the inspection completion, the training currency. The OSHA 300/301 recordkeeping is the data source in the US; the UKs RIDDOR and the employers record are the pair. Power BI DAX is the standard analytics layer; the guide below is the HSE-specific build, from the register to the scheduled dashboard.

Common mistakes

1. Plotting the counts without the rate — 7 incidents in a year means nothing without the hours and the headcount; the TRIR is the comparable number
2. One dashboard for everything — the executive view (the trend, the KPIs), the operational view (the by-type, the by-department, the open actions) and the investigation view (the detail) are three dashboards
3. No date table — the DAX time intelligence (the YTD, the same period last year) needs the date table; without it, the trend is hand-rolled and wrong
4. Publishing without the refresh — the dashboard that shows last months data is worse than no dashboard; the schedule and the data-source check are part of the build