6 Things I Wish I Had Known A Year Ago – For The Self-Taught Data Scientist in Particular and The Self-Learner in General

August of 2017 marked a year since I got a job. A year in this adulting space. A year of being a practicing Data Scientist. It is also roughly 2 years since I started my Data Science journey. Two years since, before going to the final year of my undergrad, I decided I would like to do my final year project in the Data Science space and started doing online courses and competitions to gain the skills.

Two years is a long time to have been actively learning and working towards gaining data science skills. I do not claim to be an expert. If there’s anything the past one to two years has taught me, it is the fact that the road is long and hard. Two years is nothing compared to the journey ahead. I do however have some experience and with the wisdom of hindsight, there are several things I wish I had known sooner, things I am going to share in this post, in the hopes that someone else’s journey is made somewhat smoother. Tips for the self learner in general and some for the data scientist in particular.

So here’s a couple of things I wish I had known 2 years ago:

1. …that finding my tribe would catalyze my growth

You cannot do it on your own, so find a community of like-minded individuals. For me, seeking out the Nairobi Women in Machine Learning and Data Science community and then eventually becoming one of the organizers has been amazing. I have met individuals who have challenged me to grow through collaboration on small projects, through collaboratively reading and discussing academic papers as well as just learning about how different people are applying Data Science and Machine Learning in various fields.
It has opened the door to so many opportunities for me! Through taking the time to organise the community, sometimes I find that my profile is picked up by the most unlikely individuals and organisations with a wide range of requests from authoring blog posts to inviting me to events halfway across the world and then proceeding to pay my airfare so that I can attend. (I’m editing this blog post in Douala, Cameroon! *happy dance* Who ever thought I would find myself in Cameroon!)

I will also not discount the simple fact of feeling like you belong. For a long time at work I would talk about Data Science concepts I was struggling with and have people look at me like I was just making up big fancy words to sound smart, because I am the only one who works on data full-time, but with finding a community, knowing who to approach with my struggles and learning that my troubles are not unique in some way validates my path, lets me know that I am doing something right.

I wish I had known that finding my tribe would catalyze my growth. Find your tribe!

2. …that it would take longer than a year

Like I mentioned, I have been practicing for a year and still feel like I have only scratched the surface. Looking back to a year ago, I expected I would be a data science/machine learning virtuoso by now, or something to that effect. I am not, but I can appreciate that I have come a mighty long way even though the journey ahead is possibly longer than the portion that I have covered. So I am buckling down, continuing to trust that putting in the hours will get me there and getting on with it.
Besides, I know for a fact that my journey is teaching me a greater lesson than the one I have set out to learn. So take a deep breathe, patience. Try to enjoy the process. You will get there. Just don’t stop.

3. …that impostor syndrome is a thing

…so deal with it. Ever since I set foot in the workplace, I have felt like I do not measure up. I have been the definition of saying yes to doing stuff and then learning on the job. I have been approached with problems I did not know how to solve and been challenged to learn how to, and I sometimes feel like a fraud, like I will wake up one day, go to the office and be summoned to a meeting room and informed that I have been discovered as a fraud. A year later, I am still here. I have done a lot in that year but there are still things I don’t know. Things I will be asked to do and will have to first learn how before I can deliver. I have learnt to acknowledge that feeling of inadequacy, which is sometimes irrational, and to let it challenge me. Let it be a sign that it is time to stretch even further. I am also learning to be kind to myself, acknowledging that I have done a lot in my 1 year in the workplace and that those things count.
I wish I had known that impostor syndrome would be a hallmark of my experience and had right from the start let it propel me forward as opposed to making me cower and step back.

4. …that it does not matter if they know that I do not know

Don’t be afraid to ask. So what if they know that you don’t know. For the longest time I was wildly intimidated by the fact that I work at a tech company and the field of technology comes with a lot of specialized jargon. I would sit in on meetings and fill up a whole page of my notebook listing down things which I did not know and wanted to go and Google later. I am now more comfortable just asking someone to explain things to me. You cannot be taught everything, but there are definitely instances where having someone explain things to you makes for much easier and better understanding of concepts. So just ask.

5. …that some of my models would perform no better than guess work and this would in no way be a reflection of my abilities.

I remember building my first couple of models ‘in the real world’. In the real world meaning the data wasn’t from a competition, already carefully curated and cleaned and with guidelines or hints on what algorithms to use or what questions to ask. It was all me from scratch. Getting the data, cleaning the data, engineering some features, deciding what questions to ask and what algorithms to try out. It was all me and my first several models sucked. I dwelt on them, took their poor performance personally and doubted my abilities in a big way while the reality is that some models are just rubbish. Not because of you as a data scientist but because whatever question you are asking cannot be answered by the data at hand. My advice is to find and build a pipeline or process for working through problems that you are comfortable with that allows you to build, test and iterate your models quickly. That way you rule out all the things that don’t work very early on and spend more time refining models that actually work.

6. …that I would never feel ready enough

If you are venturing into Data Science and Machine Learning on your own, or really any other field on your own, chances are you start with a couple of online courses. The field itself is wide, as is the required skill set. You will likely do one course, after which you will then start another which focuses on something different, and then another, and possibly be in this loop for a long time. The truth is you will never truly feel ready. There will always be another course that you think will supplement your knowledge and skills, always! Do not get stuck in that loop. As soon as you can, I say two courses in, introduce competitions and challenges to your learning cycle, whether they be on Kaggle, DrivenData, KDNuggets…wherever. Start doing actual Data Science work on your own as soon as you can.
I hope my little nuggets make your journey easier, give you the psyche to ‘fight’ another day.

When all is said and done, we have to know what we become. So keep at it.

To keep up with my musings, subscribe!
Like it? Share it!Share on FacebookShare on Google+Tweet about this on TwitterShare on LinkedIn


Add Yours
  1. 5

    Hi Kathleen,
    I really enjoyed your comments on the AI.Ke fireside chat recently. I’m giving a talk tomorrow in Kigali to aspiring data scientists from all over the continent. If you don’t mind, I’d like to reference your article and talk about some of the points you mention here.
    All the best,

Leave a Reply

Your email address will not be published. Required fields are marked *