Paragraph by paragraph
I am writing to apply for the Senior Data Analyst role in OVO's forecasting team in Bristol. I have worked as an analyst since 2020, first for Avon and Somerset Police and since January 2022 at Tesco, where I work on forecasting and pricing for store operations.
The team she wants to join, and her route there
She names the team as well as the company, which shows she read the advert. Her background is two employers and a start date, so the reader doesn't have to do any arithmetic.
The project I am proudest of is a demand forecasting model, written in Python, that now runs for 150 stores and has cut stock waste by about £2.4m a year. The model took around two months to build. Persuading store managers to use its numbers took most of the following year, much of it spent in their stores going through the forecasts with them.
One project, with the hard part included
The model and its result come first, then the part that is usually left out: getting people to trust it. Hiring managers for senior roles look for exactly this, because a model nobody uses is worth nothing.
I have also built 15 Tableau dashboards that senior managers use to follow performance across six business units. I automated our weekly reporting in SQL and Python too, which gives the team back about 12 hours a week. My pricing A/B tests have identified around £800,000 in possible extra revenue, which the commercial team is now testing in more stores.
More results, worded carefully
The dashboards and the automation show she does more than modelling, and the hours saved put a value on it. The pricing figure is described as identified and possible, which matches what the work actually did.
At the police I analysed crime patterns to help decide where officers were sent. Mistakes there had real consequences, so I got into the habit of checking every figure twice before anyone else saw it, and I still do.
The police years, and a habit from them
The police role adds a public sector setting and a habit that matters in any analytics team: checking figures before they go out. She describes the habit and lets the reader draw their own conclusion.
I have a BSc in Mathematics with Statistics from the University of Bristol. Before we speak, I can send you a cleaned up version of the forecasting notebook, with Tesco's data swapped for public data.
Degree and a practical offer
Her degree is relevant, so it gets a line. The offer of a notebook with public data in place of Tesco's shows skill and good judgment about confidential data at the same time.
Stating impact without overclaiming
Analysts are rarely the only reason a number moved, and experienced managers know that. The wording you choose tells them whether you know it too.
| What happened | Honest wording | Wording that invites hard questions |
|---|---|---|
| Your model is in use and the saving is measured | "has cut stock waste by about £2.4m a year" | "saved the company millions" |
| Your test found an opportunity that isn't realised yet | "identified around £800,000 in possible extra revenue" | "generated £800,000 in revenue" |
| You built a tool other people use | "dashboards senior managers use to follow performance" | "drove data driven decision making" |
| You saved time | "gives the team back about 12 hours a week" | "massively improved efficiency" |
If you don't know how a saving was measured, say what you do know, such as the size of the problem your work addressed. Interviewers for analyst roles will nearly always ask how a figure was calculated.
A shorter answer for the application form
Many analytics jobs are applied for through a form with a short answer box. This version fits a typical limit and keeps the two biggest results.
What to put in an analyst portfolio
A short portfolio helps, especially if your best work sits behind a company firewall. Two or three pieces are enough.
- A notebook that answers a real question, written so someone else could follow it, using public data. UK government data, from the ONS or data.gov.uk, is a good source.
- A dashboard built in a public version of Tableau or Power BI, with a short note on who it is for and what decision it supports.
- A short write up of a project you can't share, describing the problem, your method and the result, with nothing confidential in it.
- Clean SQL for a realistic query, with comments, if the role is heavy on databases.
One clear project explained well beats ten half finished ones. Put the link in the header of your letter, written out in full.
If you are applying for a different analyst role
Graduate or junior analyst
Lead with a project you did end to end, even a university one, and the tools you used. Show you can explain a result to someone who doesn’t work with data.
For my dissertation I analysed five years of bus punctuality data for the West of England and presented the findings to the regional transport team.
Moving from Excel reporting to analytics
Show the step up: something you automated, a question you answered rather than a report you ran, and the coding you have learned.
For three years I produced our weekly sales reports in Excel, and last year I rebuilt them in SQL and Python so that they now run on their own.
Data scientist role
Focus on models in production, how you evaluated them and what changed because of them.
The forecasting model I built now runs every night for 150 stores, and keeping it running in production is the work I want to do full time.
Public sector analyst
Public bodies care about method, transparency and explaining results to non specialists. Mention any work published or presented to a board.
I spent two years analysing crime data for a police force and briefing senior officers every month, so I am used to numbers that decide where limited resources go.
Business intelligence or dashboard role
Show who uses your dashboards, how often and what they decide with them.
The 15 dashboards I built are opened by senior managers every Monday morning to decide where store staff should be the following week.
Data you can never share
Analysts handle a lot of personal and commercial data. UK data protection law covers the personal data, and your employment contract usually covers the commercial figures. A cover letter or portfolio is not the place to test either.
- Personal data about customers, staff or members of the public, even if you think it is anonymised.
- Anything from public services such as police, health or council data, which is often especially sensitive.
- Commercial figures your employer hasn't published, like margins, supplier prices or store level sales.
- Screenshots of internal dashboards, which often contain more than you notice.
Describe datasets by size and type instead, such as "five million records" or "transaction data for 150 stores", the way Amara's CV does.
Five data analyst cover letter mistakes
Listing tools with nothing built
"Python, SQL, R, Tableau, Power BI, Excel" tells the reader what you have installed, not what you have done.
Mention each tool inside a sentence about something you built with it.
Claiming revenue you only identified
An analyst who says "generated £800,000" for an opportunity that hasn’t been realised loses credibility with exactly the people who are hiring.
Use precise words: identified, forecast, saved, measured.
Leaving out the people side
Analysis only matters if someone acts on it, and senior roles are judged on that.
Mention who used your work and how you persuaded them.
Sharing real data
Attaching real figures or screenshots shows poor judgment about the thing analysts are trusted with most.
Use public data in your portfolio and describe real datasets by size and type.
Jargon for its own sake
Model names and statistical terms the advert didn’t ask for make a letter harder to read.
Name the method only when it matters to the result, and explain it in plain words.
Numbers, units and links
- Length: about 250 to 300 words, one page.
- Numbers: consistent units (£2.4m, £800,000) and an honest word for each: saved, identified, forecast.
- Links: one portfolio or GitHub link in the header, written out in full.
- The person to address: the analytics lead, if the advert names one, with Yours sincerely. They often run the technical interview too.
The main points from this page
- 1Name the tool, the scale and the result in one sentence.
- 2Include the hard part: getting people to use your work.
- 3Use the honest word for every figure: saved, identified, forecast.
- 4Build your portfolio on public data.
- 5Never share personal data, public service data or unpublished figures.
Frequently asked questions
What should a data analyst cover letter include?
The role, your current work, one project with the tool, the scale and the result, how people used your work, and a link to a portfolio built on public data.
Should I list all my technical skills?
No. Your CV lists them. In the letter, mention a tool only inside a sentence about something you built with it.
How do I write a data analyst cover letter with no experience?
Use one complete project, from a course, a dissertation or your own interest, with real public data. Explain the question, your method and what you found, in plain words.
Should I include a portfolio link?
Yes, one link to two or three good pieces. Use public data and explain who each piece is for.
How do I talk about work I can’t share?
Describe the problem, your method and the result in general terms, with datasets described by size and type. Offer a version of the work rebuilt on public data.