Data School training reflective blog - DS24

by Liu Zhang
Photo by Vasily Koloda / Unsplash


For people who want to apply (or even accepted) DO attend as much Thursday public training (TTT) as possible. I succeed on my second application, but given the Covid time-gap, it was almost 6 months of time in between. Even though the failure on the first application was extremely disappointing, I continued to attend the TTT sessions. The skills I learnt about Tableau and Alteryx gave me a head start on the DS training, it ease the transition into the actual training.

If possible, attend Data School open day events as well even if you are not apply for the current cohort. It is always good to get your name outer there, we do remember you. (I attended two myself).

First start of DS:

Impostor Syndrome - most people have it

and certainly the biggest worry for me as well. Andy mentioned that in the very first session with us. It depends on how you take it, I took it as a drive to work extra hard, extra MakeOverMonday viz, Alteryx challenges, blogs, as I was always worried of fail to meet the requirement.

No fear is sign of arrogance, too much fear is sign of cowardliness.

Find your own level of comfort that suits you best. Do speak to people, which brings onto the next point.


Given the Covid restriction, my cohort is the first one to be allow back in person training. I would strongly recommend people to come in if they can, simply meeting more people is a good reason already, you never know when you want to ask help from someone or ask about their training/placement experience. It gives a much quicker way to learn about the company operation and future career planning. As most people in TIL has been through DS training, you are much likely to get answer if you have spoken to people. It is always good to attach a face to name.

P.S.1: Join coffee chat, get weekly assigned to a random person to talk to is great way to start meet more people.

P.S.2: There is a table tennis table, which I love to play. So I want more players to be in to practice 😁

In-person vs Zoom training:

I would strongly vote for in-person training, as you don't get enough interaction over zoom, plus only one people can speak at a time which highly restrict discussion and participation level.

Of course in some cases where the trainee/trainer is not available in London, then online session is unavoidable. But that is not an excuse for other who can attend to give a pass, it is so much easier to answer questions, debug workflows when sitting next to each other rather than talk over Zoom. The speed of problem solving in person is unparalleled, that also translate to the speed of learning.

Learn from your mistake:

Embrace the error messages.

The trainers will teach you about 80% of the stuff you need to know, that last 20% are from resolving those error messages. Tableau LOD/Table calculations are always the tricky one, while Alteryx has so many more pitfalls. I wouldn't mind stuck on one for 30 minutes or even 1 hour (as much as the frustration is miserable), because once I resolved, I do think I will remember it better.

Teaching is the best way of learning.

I thoroughly enjoy when someone in the session encountered an error, as trying to resolve it truly test my understanding of the topic/tool(s). So learn to ask for help and learn to answer for help.

Client Project:

In the second half of the training, you will receive 8 client projects. Some are external and some are internal depends on the availability. Quite often the data supplied are not clean/complete/correct, so good luck with that. It is real life experience, there is a saying

data prep tasks occupies 70% of a data scientist's time

When Andy tells you the importance of planning, I can confidently say your plan will be wrong. People always underestimate the time needed to clean/understand the data, and overestimate the time they have for dashboard building.

I have been to more than a few client presentation, where the clients simply said the number is wrong (as people didn't check duplication in the first place or after join in tables) and there were a lot awkward moments. So I recommend the team to have at least one person to be responsible for the data, just like a corporate environment where you have at least one number guy you can always ask about questions.

Work in teams not in silos

It is easier for the project manager to check overlaps or mistakes when everyone is communicating. There are people who likes to work alone, but good luck of finding examples for your interview questions about teamwork. If you think someone is slowing you down, remember what mentioned above? Teaching is the best way of learning.

For the dashboards, don't worry about the overlap when the requirement is few. There is nothing wrong to produce different ways of presenting the same information for the client to choice. Also don't produce 10 dashboards just for the sake of producing something, if they don't answer the questions asked, they are useless.


Quite often the client projects are passed onto the clients at the end of the week, so a good documentation is extremely useful. If you don't want to get constantly asked about those calculated fields, write comments. There are often the cases where the client really liked the project, so they continue to engage with TIL to expand further on the work to production level and someone from TIL will pick those up, so for the sake of future you, write comments.

I have some simple test for the quality of document your work.

1: You just drop your work now, someone else coming to pick it up immediately and they have no idea where to start. You failed.

2: Don't touch/look of your work, then you come back in one month time and if don't understand your calculation/workflow. You failed.

Get into a good habit when in training, it will make your life easier in the future when there are more stakes in the job.

Extra study and practices:

Currently the advice is to stop work at 5pm, this is not an excuse of not finishing your work, but it is advised for the work-life balance. I do agree with the ethos

Work hard and play hard

There are allocated time to do MakeOverMonday viz each Monday, but there are a lot more practices out there with Alteryx and Tableau Prep. This is a voluntary imitative for each individual to take, therefore do you want to aim to achieve 100% of what is asked of you or you are aiming higher?

As noted above, I was worried about Impostor Syndrome, so I did more work. Even though now I am much more relaxed about worrying to much, I have settled into a habit of doing more Alteryx weekly challenges or writing blogs. I treated Alteryx challenges as a nerdy version of Sudoku rather than extra work, it is more relaxation than stress for me.

While for the blogs? Shamelessly self-promoting: Blog Index

P.S. Write like you have time traveled, write an answer for yourself who was searching for answers in the past.

There are also additional service TIL provide, such as Tableau server, Snowflake, AWS. For your future career development, do you need to learn them, yes. Remember the point mentioned at the very top?

Start early, it helps.


You were expected to finish Tableau Desktop Associate and Alteryx Designer Core during the training and you will be offered revision sessions before that. If you have taken the imitative as mentioned, they shouldn't be too much of worry. On top of that, I managed to pass Alteryx Designer Advanced exam as well (which is something I am quite happy with). My advice on them: link

I was thinking of taking Tableau Desktop Professional certification before finishing DS. We had the dashboard week in the final week of training and I was waiting for that to see how ready I am. (Happily Andy said I am ready) Unfortunately, Tableau is currently waiting for a transition in the certification platform change, so it is put on hold at the moment.

AWS Cloud Practitioner, something I find it is certainly feasible given they supplied with some excellent free training videos. Again taking the initiative of leaning some Udemy courses is advised.

Snowflake, not yet started, but if you have some SQL background, then it is certainly something worth try to learn as well.

Last word:

Just like playing a compute game, reaching the level cap was the end game, but when you finally did, you realise that the actual game has just started.

Now I am ready to start my first placement, am I worried? Yes. But am I as worried as I first start the Data School? No. Because I have passed my training and I am ready for the new challenge.

Looking for more guides, tips and tricks in Tableau or Alteryx? Go check out the other blog posts from the Data School.

Tableau Public



Tue 14 Dec 2021

Sun 03 Oct 2021

Tue 28 Sep 2021

Fri 01 Oct 2021