Data analyst - where to start?

[Bloodrayne]

Hi all,

I've been spending a lot of time tidying/combining data from various sources and producing stats as a part of my role; using a combination of basic SQL, PowerBI and Excel to produce various reports on things (although nothing that had to predict trends, so more reporting on historic data than things for future planning).

I've found it quite interesting, and have always had a reasonable grasp of Excel, etc - so I wondered about moving into this as a more full-time career.

However, I need to retrain and get qualified in this area (as I do not have any qualifications, and have just picked up SQL, Excel, PowerBI, etc as a consequence of wanting to expand my role from it's day-to-day routine). Can anyone provide guidance on what qualifications I should be looking at, as well as any experiences of working as an Analyst?

Thanks!

 

[HangTime]

You don't need qualifications for this in my experience (although that's not to say they don't help). I have a bunch of different data roles in my team and it's not something I've looked for specifically when hiring analysts, I am more interested in how they approach certain problems, what questions they would ask stakeholders, and of course I throw in a couple of practical questions to check they have at least a basic understanding of SQL etc.

SQL is a great base skillset to fall back on and will be in demand for many years to come but keeping in mind industry trends, if you were looking to invest in training then I would look at DS languages (Python / R), data pipelines in cloud platforms etc. Although it probably helps to think about what your career goal is, do you want to just become an analyst and take it from there, or do you envisage focusing more on tech (architecture etc) or stay heavily business focussed?
Also consider if you would like to become more of a Technical Business Analyst, and if so having some formal BA training alongside practical data experience would likely be sufficient. Hybrid Data/Business Analysts are very useful in my sector, I need people that can both elicit requirements but also translate that into what it means in data terms, do data profiling to prove/disprove hypotheses and business cases, explain why the number coming out at the end of a process is 99 when the user expects 100 etc, discuss whether it really matters if the answer is 99 or 100, etc etc.

 

[wesimmo]

As a recruiting manager looking to bring in people to junior analyst roles I've always looked at what peoples mindset is and not at what packages they can use.

You'll find you use all sorts of tools through your career, whether you use them well is down to your mindset.

I can point to at least 3 people who I noticed working in call centres, took on into junior planning analyst jobs and who are now managers themselves, simply because they were inquisitive and willing to self learn, as you seem to have demonstrated.

 

[Selekt0r]

I'd agree with everything HangTime said above. I started (quite a long time ago) doing basic 'desktop' business/data and SQL analysis when I worked in Banking, progressed into 'data warehousing', proper BI/reporting and then on to application development. Finally moved into consulting and now work as a Product Owner for a big Data/Analytics company. I have precisely zero formal training/qualifications, however I do have the aptitude and experience to recognise and understand a business problem/challenge, unpick the functional and non-functional requirements for a solution to address it (including the underlying data/data model and analytical requirements), and define and lead a sequence of work to deliver that soution.

I would definitely recommend supplementing SQL with R/Python-type skills too as these are some of the hottest and most in-demand skills at the moment.

 

[dowie]

This job title can be a bit broad tbh... two types I've seen have basically been either someone with a reasonable knowledge of statistics and some programming skills or can also mean essentially a business analyst who works with data. I guess another poster has mentioned a "planning analyst", not sure exactly what that is but perhaps similar to the business analyst type roles.

If you were interested in the more Business Analyst type roles then entry requirements aren't high (as per the replies above) though this course might be of interest (especially if you already have a degree):

https://www.conted.ox.ac.uk/about/advanced-diploma-in-data-and-systems-analysis

• The System Development Life-Cycle and Management
• Systems Analysis and Design: Data Driven and Procedural Techniques
• Systems Analysis and Design: Data Organisation
• Systems Analysis and Design: Objects and Architectures
• Team Project

On the other hand if you're more interested in the roles that overlap with data science roles then getting a relevant undergrad would be a start or if you already have a non-stats/maths undergrad then a grad certificate or diploma (can take like 1-2 years part time).

If you don't want to quit your job then The Open University, University of London Online and Birkbeck college, UoL all offer BSc degrees in Mathematics and Statistics or Data Science... or I think in the latter two cases, graduate diplomas or certificates in mathematics/statistics (sufficient to get you into an MSc in Stats/Data Science).

Also no reason why you can't switch into a relevant job part way through studying.

If you then want to progress/take it further then you can go for an MSc too - part time online or evening options available for MSc Stats at Sheffield (online) or Birkbeck (evening).

Or more geared directly towards data analyst/data science roles there are these online courses from Imperial and Dublin

https://www.ucd.ie/courses/msc-data-analytics

https://www.imperial.ac.uk/study/pg/mathematics/machine-learning/

Or if you're able to get time off during the day or can take a year out from work then a bunch of universities offer relevant taught masters programs in stats, ML, data science etc...

Lastly Udacity is very practical - they have a data analyst nanodegree, which will give some basic practical knowledge of python, SQL, basic statistics, data wrangling and data visualisation. Datacamp is also worth a look.

 

[drakulton]

Sorry, quickly read this as anal cyst, sounded a bit gruesome.

 

[dowie]

Sorry, quickly read this as anal cyst, sounded a bit gruesome.

I knew a gay quant who used to call himself a "quantitative anal-ist".

He was also a fan of colourful data visualisations.... such as his "Big Gay Swap Curve" (it was basically a rainbow!)

 

[Bloodrayne]

Thanks all - that's really useful feedback to have; for either of the roles (business or data focused) I would be tempted to get some sort of formal recognition for my training as it's only a part of my current job, and the likelihood would be that I would have to move companies to pursue a career in this area - and it probably would appear better if I have something (regardless of level) that confirms training & competence in areas - from the feedback given it generally seems that SQL, Python & R would be useful.

On the two streams of analyst, this is also very useful to know - I had not thought much more than I would like to work with data more (it's a highlight of my working day currently), data science looks very intriguing - I'd be starting from scratch when it comes to qualifications (I asked recently about whether I should resit my Maths GCSE as when I did it many years ago I did not particularly cover myself in glory), but I see that the OU does an intro to mathematics concepts that then leads into data science related qualifications. I also like the look of the Udacity course - but need to work on my Python before I can go for that.

Certainly food for thought though!

 

[wesimmo]

Thanks all - that's really useful feedback to have; for either of the roles (business or data focused) I would be tempted to get some sort of formal recognition for my training as it's only a part of my current job, and the likelihood would be that I would have to move companies to pursue a career in this area - and it probably would appear better if I have something (regardless of level) that confirms training & competence in areas - from the feedback given it generally seems that SQL, Python & R would be useful.

On the two streams of analyst, this is also very useful to know - I had not thought much more than I would like to work with data more (it's a highlight of my working day currently), data science looks very intriguing - I'd be starting from scratch when it comes to qualifications (I asked recently about whether I should resit my Maths GCSE as when I did it many years ago I did not particularly cover myself in glory), but I see that the OU does an intro to mathematics concepts that then leads into data science related qualifications. I also like the look of the Udacity course - but need to work on my Python before I can go for that.

Certainly food for thought though!

What is it about working with data that makes your day?

Is it the problem solving around collecting, collating and cleaning?

Or do you like taking data and bringing it to life, building dashboards for users, the old data vs information thing.

Is it understanding the data and identifying anomalies that indicate something has gone wrong somewhere and then investigating what and how to fix the processes that caused it?

One thing I find interesting in my field, call centres, is the way you can read human behaviours from data and you play a game of cat and mouse with agents trying to scam the system.

At the same time, it's a good mix of strategic planning, building 5 years financial forecasts at one end, and absolute pants on fire days when the systems fail and you've got hundreds or thousands of calls waiting.

 

[Bloodrayne]

Currently I take data from a few sources (internal systems, spreadsheets, databases), combine and validate these and present the findings in a clear and readable fashion with dynamic dashboards - so taking large amounts of data from various sources, ensuring that it's consistent and correct, and presenting readable/digestible summaries comes with a sense of achievement for me.

Beyond that, investigating data anomalies is of interest and certainly appeals, as does forecasting/planning based on previous data, but I've no actual experience in either so it's one thing to find things of interest, and another to see if it's something you could actually do.

 

[wesimmo]

Currently I take data from a few sources (internal systems, spreadsheets, databases), combine and validate these and present the findings in a clear and readable fashion with dynamic dashboards - so taking large amounts of data from various sources, ensuring that it's consistent and correct, and presenting readable/digestible summaries comes with a sense of achievement for me.

Beyond that, investigating data anomalies is of interest and certainly appeals, as does forecasting/planning based on previous data, but I've no actual experience in either so it's one thing to find things of interest, and another to see if it's something you could actually do.

It was an interesting discussion i was having on LinkedIn the other day about forecasting.

Since I accidentally fell into the Call Centre world, it's changed from when we very much based forecasts on historic data.

What I've found over the years is that, as businesses now try and change so much, forecasting is as much about building relationships, understanding upcoming projects and working with the business to try and understand those changes.

e.g.

They release a new product / marketing campaign, how will that impact on calls, usually optimistic views from the owner of the initiative....
There is new regulation, how does that impact on call times?
A new bit of tech is being implemented to improve processes, will that reduce call times, or contacts in the first place.

If you're more on the social side than some analysts, you might find that kind of work interesting. You still get to do the data things, but it's less stuck in a corner behind a screen pushing reports out.

 

[Bloodrayne]

That does sound interesting - I'm quite happy to be stuck behind a screen pushing out reports (it's probably a hangover from years of playing CM93/94), but developing understanding of businesses and then using that to predict potential future trends sounds fascinating too.

I guess it depends on how social the social side is, are we talking leading breakout groups/presentations? I'm comfortable when backed up by data, or with a process in mind, but less so when in more generic presentations (I prefer to think of it as I prefer to work collaboratively rather than speak at people...)

 

[dowie]

Thanks all - that's really useful feedback to have; for either of the roles (business or data focused) I would be tempted to get some sort of formal recognition for my training as it's only a part of my current job, and the likelihood would be that I would have to move companies to pursue a career in this area - and it probably would appear better if I have something (regardless of level) that confirms training & competence in areas - from the feedback given it generally seems that SQL, Python & R would be useful.

Just a note - Business Analysts aren't necessarily more business focused... it's often the other way around actually. Business Analysts are frequently (but not exclusively) under IT head count and involved in writing specs for IT related stuff, whether software or otherwise. You might also see them referred to as systems analyst too... the Oxford course would probably be ideal for this sort of role. So a data analyst that is also kind of a business analyst could also be quite IT focused and involved with writing specs, liaising with developers and with stakeholders on the business side etc..

Other data science/data analyst roles might be unrelated to IT or IT/software projects... they need to code etc.. but they might well fall under the business side itself rather than IT. For example a friend of mine models and forecasts house prices, he's got nothing to do with improving any systems, getting involved in data driven design, writing specs etc.. beyond implementing his own models.

Of course just to muddle things further - there are data science/analyst roles that are very tech focused... but not necessarily involved the business analyst/speccing stuff part... see any number of roles in big tech firms, or indeed things like detecting credit card fraud in retail banking, reporting on click data/marketing etc.. etc..

I'd be starting from scratch when it comes to qualifications (I asked recently about whether I should resit my Maths GCSE as when I did it many years ago I did not particularly cover myself in glory), but I see that the OU does an intro to mathematics concepts that then leads into data science related qualifications. I also like the look of the Udacity course - but need to work on my Python before I can go for that.

It could be a long process - beauty is though that if you enjoy it then you can just plod your way through part time.. I mean just getting some basic skills doesn't necessarily take too long, I don't think you'd need to spend too much time learning Python in order to get to the point where you can tackle say a Udacity nanodegree, I'm pretty sure they have standalone introductory courses if you search around on their website too. Something like that would seem to be a good move to just get you started/get to the point where you can do stuff with data.

There are a wide range of roles involving data, some attract average pay, some are possibly the highest paying careers/professions available.

This isn't actually anything new either - when you start to learn statistics for example you'll come across, for example, Student's t-distribution - this was created by Willam Sealy Gosset, a "data analyst"/statistician working for the Guiness brewery back in 1908! See also the insurance industry, they have made use of data analysts for centuries - they're called actuaries and they have their own set of professional examinations, this is probably the highest paid profession out there.

So basically you could get some basic skills now and try to land some sort of job doing *something* with data, perhaps you stop there and/or just end up learning new packages (business intelligence related stuff, become more proficient in SQL etc..). Alternatively, you can carry on studying, gain more academic quals... this could take the next decade.

I think UoL external system and Birkbeck will require A-Level maths, I guess the OU might be possible though starting from scratch - it could take you 4-6 years from next September to get a degree. Then if you want some of the more serious roles you'll generally need an MSc too - that's another 2 or 3 years part time, alternatively if you train to become an actuary that is at least 3 years IIRC (in theory you don't actually need a degree to become an actuary, just A-Level maths, but most are graduates IIRC).

If in 10 years time you've got 10 years of python, SQL etc... and applying statistics to business problems + BSC maths/stats/data science + MSc in similar from a good uni... you'd probably be very very employable by then.

 

[CAPTAINBUMFLUFF]

Having worked as a analyst and now as a data scientist In different companies I would say that the technical requirements differ quite a bit by company.


In the first two companies I worked for excel was the main tool used by analysts along with SQL and powerBI. I feel things are slowly moving more and more to having Python/R as a requirement.


Based on what your experience I would say you could make the jump now and apply for data analyst roles. Whilst formal qualifications are useful depending on company they might not matter so much, it’s more about what you can do. Once in you can self study and learn python. Also once in a team you can then learn off colleagues whilst at work, I.e if all the data is on GCP you will find yourself learning bits of python and about the cloud.


For data science roles however they mostly require a degree and I agree with Dowies suggestion around further study.

 

[TNGL]

See also the insurance industry, they have made use of data analysts for centuries - they're called actuaries and they have their own set of professional examinations, this is probably the highest paid profession out there.

In what I do and the companies I've worked for, we (actuaries) typically have data analysts and business analysts to create database queries, clean up data, reconcile it, create dashboards etc for us, so that we can focus more on the modelling and 'specialist' areas of the role. Don't get me wrong, we do have to clean/reconcile data ourselves but the time spent doing so is far outweighed by the time spent on the modelling, stats, commercial considerations etc.

In short, OP, if you want to work as a data analyst in an insurance company, apply to be a data analyst in an insurance company. I would not apply to be an actuary if you would like to do the work of a data analyst in an insurance company.

With that said, if you like the idea of being an actuary, I would strongly recommend it - I find the work is rewarding, it keeps me on my toes, I continue to learn a lot and it offers a good work/life balance. However, you need to be committed to the exams as these will have a major impact on your pay and progression. Also, actuarial work is very broad, spanning pensions, investment, life insurance, non-life insurance, reinsurance etc and so it's important to consider what area of actuarial work would most interest you. Within insurance there will be further segments within this; pricing, reserving and capital modelling.

I think the minimum requirement to join the institute is an A-Level in maths or equivalent. You don't necessarily need to go to university either as some employers offer apprenticeships and they're relatively well paid too.

If you have any questions, let me know.

 

[dowie]

In short, OP, if you want to work as a data analyst in an insurance company, apply to be a data analyst in an insurance company. I would not apply to be an actuary if you would like to do the work of a data analyst in an insurance company.
.

Point of the post was just that "data analyst" is a broad term... what you've described re: someone with the job title "data analyst" doing in the insurance industry is the sort of business analyst type role I mentioned, that's common enough elsewhere and is what others have mentioned too (in fact part of it (cleaning up data) of it might fall under a data engineer role in other industries too), but that isn't necessarily what all "data analysts" or "data scientists" do elsewhere.

For example one uni friend forecasts house prices, a more junior team member who works with him has the job title "data analyst" rather than data scientist... main difference is lack of relevant post grad degree... the junior guy is basically employed as a statistician too though and not as a business analyst++.

That's the point of my post, not that if he wants to be a business analyst type data analyst he should become an actuary but rather that there are (broadly) different roles that could be described as "data analyst"... and that applying stats to business problems isn't new - Guinness brewery and Gosset or indeed your industry where (that type of) data analyst/statistician is an "actuary" complete with formal qualifications and a professional body and probably one of the highest paid professions going.

 

[OspreyO]

I would say it helps enormously to be interested in the data you're analyzing. Otherwise it becomes a bit dry.

 

[TNGL]

Point of the post was just that "data analyst" is a broad term... what you've described re: someone with the job title "data analyst" doing in the insurance industry is the sort of business analyst type role I mentioned, that's common enough elsewhere and is what others have mentioned too (in fact part of it (cleaning up data) of it might fall under a data engineer role in other industries too), but that isn't necessarily what all "data analysts" or "data scientists" do elsewhere.

For example one uni friend forecasts house prices, a more junior team member who works with him has the job title "data analyst" rather than data scientist... main difference is lack of relevant post grad degree... the junior guy is basically employed as a statistician too though and not as a business analyst++.

That's the point of my post, not that if he wants to be a business analyst type data analyst he should become an actuary but rather that there are (broadly) different roles that could be described as "data analyst"... and that applying stats to business problems isn't new - Guinness brewery and Gosset or indeed your industry where (that type of) data analyst/statistician is an "actuary" complete with formal qualifications and a professional body and probably one of the highest paid professions going.

Sure, with you .

 

[Bloodrayne]

Thanks for all the feedback - I didn't realise it was such a broad term when first making the post, so this has all been very educational!

Lots to research further, but I really do appreciate the feedback given by all.

 

[XeNoNF50]

Hi all,

I've been spending a lot of time tidying/combining data from various sources and producing stats as a part of my role; using a combination of basic SQL, PowerBI and Excel to produce various reports on things (although nothing that had to predict trends, so more reporting on historic data than things for future planning).

I've found it quite interesting, and have always had a reasonable grasp of Excel, etc - so I wondered about moving into this as a more full-time career.

However, I need to retrain and get qualified in this area (as I do not have any qualifications, and have just picked up SQL, Excel, PowerBI, etc as a consequence of wanting to expand my role from it's day-to-day routine). Can anyone provide guidance on what qualifications I should be looking at, as well as any experiences of working as an Analyst?

Thanks!

You don't necessarily need formal qualifications relating to data analysis to get a job doing it.

Just my two cents, when I hire for my teams the main thing I look for, especially for a junior role, is attitude. It's always possible to learn the technical skills and the syntax of any language needed to run analysis.

I've found too many people are just processors and fetchers of data whereas at least for my purposes, to work for me, I like a curious mind, someone who demonstrates they understand the context around the data they have put together and have demonstrated they are able to use the data to make business positive impacts.