Somewhere on your council's website there is a spreadsheet describing your neighbourhood in more detail than any press release ever will: repair requests, inspection results, planning decisions, bus timings, school attendance. Nobody sends out a note telling you it is there. That is precisely why it is worth opening.
Local open data rarely hands you a headline on its own. What it gives you is a map of where to look, plus the numbers that make an anecdote stick. Here is how to find it, read it honestly, and turn it into reporting your neighbours will recognise.
Start with a question, not a dataset
Browsing a data portal is a reliable way to lose an afternoon. Begin somewhere else: with something you have already noticed. A crossing that feels dangerous. A street where bins are collected late. A clinic that keeps closing early. Write the question down in plain language, then ask what record would show the same thing at scale.
Workable questions are specific and bounded. "Are street repairs in this ward taking longer than they did three years ago?" can be answered with dates and locations. "Is the council failing residents?" cannot. If your question cannot be met with a count, a rate, or a comparison across time or place, sharpen it before you download anything.
Where local data actually lives
Open data is scattered across more places than most people expect. Check these before concluding that nothing exists:
- City, county or municipal open data portals, usually with a dataset catalogue and a search box.
- National statistics agencies, which often publish figures down to district or neighbourhood level.
- Budget and spending files attached to council papers, sometimes as PDFs you will need to convert.
- Inspection and licensing registers: food hygiene, building permits, taxi licences, environmental monitoring.
- Transport authorities: timetables, punctuality records, roadworks, cycle counts.
- Public request archives, where other people's freedom-of-information requests and the replies are published.
- Community sources that are not officially "open data" but are still data: a residents' group spreadsheet, a school catchment map, readings from a local air-quality sensor.
If the file you want does not exist, that is itself a finding, and often the start of a request. Ask for the record, not the answer. "A list of repairs completed in the last two years, with dates and locations" gets you further than "why are repairs so slow?".
Read the metadata before you read the numbers
Every dataset was collected by someone, for a purpose that probably was not journalism. The metadata — the notes that come with the file — tells you what the numbers can and cannot support. Skipping it is how careful reporters end up with corrections.
Four questions to ask of any dataset
- What is one row? A complaint, a building, a day, a postcode? If rows are incidents, you cannot say how many households were affected.
- What is the unit, and is it a total or a rate? Counts rise with population. Rates let you compare a busy district with a quiet one.
- Who is missing? Records logged by phone only, or by one department only, describe a slice of reality, not all of it.
- What period does it cover, and when was it last updated? A file refreshed quarterly may be months behind the complaint sitting in your inbox.
The traps that catch careful reporters
Boundaries do not line up
Health, policing, schools and councils all draw their own maps. A dataset for one area will not match another, and joining them produces nonsense. Always note the geography a figure belongs to, and say so in the copy.
Missing data is not zero
A blank cell can mean nothing happened, nobody recorded it, or the record was withheld. Treat blank, unknown and zero as three different things until someone confirms otherwise.
Changes in method look like changes in the world
If a council switches systems or redefines a category, the line on your chart jumps without anything real happening. Ask whether definitions changed. A short note in the methodology is worth more than a dramatic graph.
Small numbers and privacy
Small counts can identify individuals, especially alongside addresses or dates. Aggregating, or leaving the smallest categories out, is usually the right call. Rules on what authorities must release, and on what you may publish, differ by country; if a story turns on either, run it past an editor or a lawyer before it goes out.
From number to story: three angles that work
Change over time. Show the same measure across several years. Rising, falling, or flat — each raises a question a press officer has to answer.
Comparison within a place. The gap between the best and worst district in your coverage area is often the story, provided the underlying data is collected the same way in both.
The promise against the record. Budgets, strategies and manifestos set expectations. Spending files, inspection logs and completion dates show what happened. The distance between the two is your article.
Whichever angle you take, pair it with one concrete case. "Repairs averaged six weeks" is abstract. A named street where a broken light stayed dark through winter is what readers remember, and the two belong in the same piece.
Put people back into the numbers
Data tells you where to stand; it does not tell you what it feels like there. Call the residents behind the complaint records. Ask the official whose department logged the entries to explain the process, and give them room to respond to what you found. Send your working — the file name, the date you downloaded it, the filters you applied — so the argument stays about the evidence.
Then show your readers the limits. Say what the data covers, what it leaves out, and what you could not confirm. Admitting a gap builds more trust than a confident overreach, and it protects the story from the first person who opens the same spreadsheet.
A first week you can actually manage
- Pick one question from something you have noticed on your patch. Write it in one sentence.
- Spend an hour locating one dataset that speaks to it. Download the file and save the original.
- Read the metadata, then check five rows against the source document or a phone call.
- Call two people: one affected resident, one official who can explain the process.
- Keep a simple log of every file, date and filter, so the reporting is reproducible.
Local open data rewards patience more than brilliance. The council publishes it because it has to, not because it wants to be read. Read it anyway, one question at a time, and you will know your area better than most of the people paid to describe it.
Photo: ohalso / Pixabay
