To ask you what your day looks like if you work as a data analyst?

[sink7]

I am a biotechnology scientist and am used to generating/analysing data, making lovely graphs, etc.

I would like a better work/life balance and the ability to wfh and thought data analysis might suit. However, when I look at job postings I am pretty thoroughly baffled by the vague language and struggling to picture what a data analyst might actually do on a day to day basis.

Any insights appreciated!

I work with some data analysts for a large company. A typical day is
Troubleshooting data discrepancies (reports sent out to users who will say something doesn't look right, generally down to how someone has processed something in the system).
Regular reporting - preparing monthly performance reports, which involves getting data from legacy systems and converting it into formats that people can actually use. Then meeting with management to talk through the overview of the reports - any issues with data that may need to be taken into account in the results, any future changes they may want to make to the reports, and a highlight of the reports e.g. any anomalies to expected figures
Project work- this is ad hoc with different teams. Liaising with the ultimate business users (so need a good knowledge of how the business operates and what they are trying to achieve or visualise in the context of business targets) and IT (who don't have much of a clue about the business but know the legacy systems inside out and how to extract the data from it) to formulate different reports and visualising data in different systems. There is lots of iterations of these reports and lots of back and forth and their role is to explain what's going on to both sides as IT and the end business users can't often understand each other. They also question the data and how its visualised, to make sure it shows users what they actually think it is showing them. So lots of meetings
Also ad hoc one off data requests (someone will ask 'how much of x business do we have and how much profit does it make) and there isnt an easy way of getting that info...this probably isn't very interesting for them as there isn't much data interrogation or manipulation or actual analysis

They are in the office one day a week, occasionally 2

And they have to do a qualification in data analysis

Personally it sounds similar to the work you've already done.
I run reports to gather data from the previous day. I check it to validate the updates people have done are working correctly. Checking for anomalies, looking for gaps or problems so they can be fixed. We also run historical reports as requested by other teams and use the data to answer questions they have.
I make graphs to illustrate our work so our metrics can be shown to directors. Lots of work with SQL, Excel and similar.

In my team though there are four of us with analyst in our job description and we all do quite different work.

What PP said.... As I've progressed I've moved away from regular report creation and more towards project work which is more interesting, but still spend a lot of time looking at discrepancies and getting to the bottom of issues

Have you looked at roles in the analytic function within the civil service more generally?

There are loads of different roles in all sorts of departments.

From observing DP while WFH, largely endless meetings, some coding, wandering round the house making endless cups of tea. Lots of flexibility and autonomy. Waiting for numbers to crunch then presenting them to people then more meetings. More cups of tea.

Proper mix as others have said:

• regular reporting on KPIs, Statutory requirements - creating/updating dashboards
• project work, data analysis to support direction of travel, decision making, benefits realisation
• many many meetings to present the above plus advice/guidance on new initiatives
• data wrangling to get said data into useful usable format (far too much data cleaning)
• working within the data governance framework to bring about data quality improvements

I would say about 50% of the time is spent doing data analysis (less if at senior level) and 50% doing all the other gubbons that comes with it.

I'm a data analyst and my day looks similar to what stuntbubbles said.

At the beginning of my career my day was very much task-based - so largely small-scale tasks that could be accomplished in a few hours/days. These would be allocated to me by my manager. This used to be stuff like finding out sum/min/max/average etc of values over a period of time and then plotting it in a graph or extracting data into excel spreadsheets.

I then started being given projects and taking more ownership over my work day. At this point, I was involved in more meetings - firstly as an observer and then (eventually) as a 'specialist'. My role in meetings was largely to advise on what data could be used to help answer people's questions/problems, agree on a timeline for delivery, and then present updates and results. I worked largely as a bridge between the business and the data team lead in this respect and had a larger role in the more technical elements of what the data team were doing (e.g. helping ingest data sources into our database, building visualisations for business operational use on Tableau, identifying solutions to existing business problems and writing reports on how they could be implemented - getting signoff and then actually completing the projects end to end).

Now I'm 5 years in and I've asked to not be involved in so many meetings (because I've found them to be a waste of time). So essentially I'll liaise directly with project managers in different areas in the business and get them to provide me a summary of the meetings I used to be part of. I then propose solutions or projects to help leverage the data to help them, get their agreement as to whether to go ahead, get agreement from the data team lead, and then complete or manage whatever needs to be done end to end).

A typical day is as follows (I now work from home):

• roll out of bed around 9am
• make myself a coffee
• catch up on emails/messages/check the requests forum at work to see if there's anything urgent to acknowledge/attend to
• continue working on projects/tasks that I have outstanding from the previous day (this'll include stuff like extracting data, creating new tables/views/schemas in our database), building data visualisations, writing reports)
• lunch and catching up on YouTube
• meetings
• send updates to my stakeholders on the status of the projects I'm working with
• help other members of my data team with the tasks they're struggling with
• report on my progress to my manager

In terms of projects and tasks I've been involved with, it's been a bit of everything but that's because I've always chosen to work at startups and always jumped on anything that looks challenging/new/different. So this is a summary of some things I've worked on in the last 5 years (I've tried to keep it chronological):

• building powerpoint presentations that summarise research or analysis that's already been conducted/findings from current projects
• identifying data sources and systems to ingest into the main database/data lake
• extracting and cleaning data from the database/data lake using SQL
• using Excel to create pivot charts and simple visualisations for business to use and adapt for their needs
• visualising data on software such as Tableau, PowerBI, MicroStrategy, Amazon Quicksight, Metabase etc (whatever the client wants!)
• creating tables, views, schemas, pipelines in the main database (SQL)
• taking on Machine Learning tasks (such as prediction models and text analytics) using Python

In terms of work day, there's definitely organisations where you'll only be working 9am-5pm without needing to put in extra hours. However, if there's an urgent task then you may occasionally be working into the evening at times. At the moment I have some days where I do practically nothing all day, but these are balanced out with the days where I'm working 12 hours! I'm not the strictest person with regards to time management though - if I wanted a perfect life balance then I could achieve it, however I enjoy procrastinating!

Overall, it's the best "office job" I've done - easy money for something I don't find difficult. Best of luck OP, sounds like you have the right skills for it!

How does one become a data analyst? It sounds like I have some transferable skills, especially the bits I enjoy most about my job.

Thank you, this is very helpful.

Is a lot of the manipulation manual, or do you rely pretty heavily on scripts to transform data? My coding knowledge is essentially nil, so this is where I would really struggle.

@pippinsleftleg I did a MRes previously (which covered Quantitative Research Methods) which I used to land my first data science internship...which was followed by another internship (because I moved country) which translated into a full-time job.

For the first internship I had to demonstrate that I had the capability to learn SQL before the internship commenced, but they were happy with the fact I'd already done some predictive modeling as part of my uni course (on SPSS). I had no previous coding or programming knowledge. I used w3schools to learn SQL in a few weeks.

Nowadays I advise people to do an online data analysis course (e.g. through coursera/datacamp etc) so they have something to put on their cv and demonstrate their commitment. Perhaps also have a portfolio project linked on their LinkedIn (or even a github repo) and then get applying!

@sink7 "Is a lot of the manipulation manual, or do you rely pretty heavily on scripts to transform data?"

• most of it is scripted through SQL during the data extract, because if a request comes in once it's likely to come in again! So it's better to script it in SQL and then apply finishing touches (if required) in Excel if absolutely required.

However, that having been said, it's not like you have a "data cleaning script" that you apply to everything - you've got to play around with the data and take whatever steps you need for that particular dataset to get it into the right shape needed for that project.

If there's a database table that has known issues that would need the same cleaning steps taken irrespective of project, then you could either create a new 'cleaned' version of the table or create a view which you could use to extract data in future rather than constantly applying the same steps.

I promise you both that SQL is not difficult to learn though. The main challenge is being able to identify what your stakeholders want from the data and then manipulating the data in the right way to provide an accurate answer

I went from data analysis into statistics, but still quite similar. I've worked across a lot of industries, and they are all a bit different to each other but as you get more experience you can apply skills to everything. Depending on your department/organisation, funding of technology differs wildly - some places have been heavy on coding, some jobs have been entirely in Excel. One wanted me to print out pages of numbers and manually count 😳 but I swiftly put a stop to that! Look at the job ads and ones that day Excel/Word/PowerPoint are probably less code based than those asking for SQL, R, Qlik etc.

I analyse data all day. Sometimes for variety I look at data and analyse it.

@CaribouCarafe your last line there reminded me of a tweet I saw - in order to replace data analysts with AI you'd need clients to explain clearly exactly what they want, so our jobs are safe 😂

@CaribouCarafe Thank you for such a comprehensive reply! A large part of my job is presenting data to inform business decisions, so I am comfortable with that aspect of things. I am a bit concerned that even entry level jobs seem to ask for candidates to be confident coders.

How did you get started, if you don't mind me asking? Did you do a degree in maths/stats, or did you transition from a related field?

Asked and answered in a previous comment - thanks!

Thanks all for your very helpful replies. I am pretty confident in my ability to analyse data within my current field, but worried I would be hugely out of my depth in any other environment. 😅

I am trying to learn a little bit of python via datacamp and am keeping eyes open for apprenticeships/entry level jobs.

@CaribouCarafe thank you - it’s something that I’ve thought of before but didn’t really know where to start - I’ll start looking at courses

@sink7 there's a mix of skills in the team. Some use SQL, some Excel or R but actually we've started using Alteryx quite heavily and that does the job very easily without the need to learn loads of coding. We're also Tableau users so the two work together nicely.

Why not have a look at somewhere like the Information Lab. They do loads of free sessions in tools like these. Might also be worth checking out their Data School if you're serious about a career change.

This made me laugh! So true!

There are free online courses to learn basics of SQL.

I want to suggest you look at BI, too. I've worked in Business Intelligence (BI) and research science. I sometimes look at data analyst jobs. There's overlap with them all, I suppose. Our BI work environment was somewhat under-resourced, very structured, very managed, not lazy coffees & pondering data: we had plenty deadlines, regular catchups, strictly managed holiday entitlement, people to CC when reports were ready, etc. We had key clients whose requests took priority over BAU work.

90-95% of what people describe here is what we did in BI. Turning numbers into pictures. Except that it was very corporate, dress smartly, be in office, and our reports were much more routinely generated rather than implying creative data incorporation elements (we had SOPs to maintain). Tableau or PowerBI rather than R although some people in a related team used Stata. I read that key difference BI is focused on making sense of what just happened while data analysts try to predict the future. Actually, in BI, we sometimes did trajectories, too, just not with very sophisticated methods.

Because I worked as a research scientist but then I have had to explain why data scientists were wrong in doing something similar I would say... on average, the data analysts are the most naive about their data. Over-confident Again and again. The BI people know the data are inaccurate and call everything "indicative". The research scientists spend ages writing up how biased the data are so how limited any conclusions can be. The data analytics people, that I encounter, are way too under-concerned about data quality and bias. Maybe they think that decisions have to be made & information quality just doesn't matter as much as quick decisions.

I do think the data scientists at their best do produce the most robust measurements of things, they spend a lot of time thinking about what the most reliable & consistent interpretations and uses of datasets can be, how to make the (inherently always inaccurate) data be as useful as possible. Guidance then used by the BI people. I have immense respect for the people who work for the ONS, for instance.

Think about other research type jobs as well (market research or social research) I started as a data analyst for an independent research company, then moved to managing research projects for clients before moving into the public sector to handle research from the client side. If you're talented with numbers, can problem solve and then also communicate your results clearly you're a rare breed as many people can do one or the other.