black pendant lamp

OMG, I did it ya’ll I MOFO completed all the levels of York online CS master. I finally handed in my final dissertation after 0 sleep and boom that was it, I was done. No more guilt. No more deadline stress. No more living only 10% of my best life. Basically I wrote this blog post 4 months ago but hadn’t finished it so the rest of this post will sound like I had just completed it! I am finally a FREE woman! Dances.

It’s exciting getting to the IRP stage because you get to choose your own project in your own field and find your own two feet. Finally, a module where you are not rushing through content and reading 1000 pages of a book about how to calculate disk rotations. I learnt so much but as per usual I struggled majorly, with writing, in this module but now I know I have dyslexia I finally understand why anything academic-related is a killer for me. Heh. If I can do it – so can you all! I am going to try and collate some advice and tips to stop you all from making all the same mistakes as I did.

First of all this module is worth double the credits and takes 4 months instead of 2. The MOOC (online content) will only contain 4 β€˜sections’ one for each month filled with things you will need to cover. But honestly, you can get through those 4 sections in about a week. There’s not much reading just references to small bits of advice, links and reminders of what you require to build your dissertation and conduct the research. They will recommend you to actually complete the first month’s section in the first week, 2nd section about the second month in the second week etc etc but I would rush through it to see what it’s all about and know where bits of information are – then reference back to those bookmarks when you start working on those bits.

My Experience

Your final project is essentially within YOUR power to do anything you possibly want in your field of choice.

So, me being me decided to do the most complicated thing possible which was to choose a topic we have NOT even covered in any module and that I didn’t have any personal experience with either πŸ˜³πŸ˜‚πŸ˜­πŸ˜­ I went into the Artificial Intelligence pathway and basically decided to write a Recommendation Model – FML.

Although it was the most basic ML method in the field it was pretty damn difficult to learn it from scratch. But I wanted to build my idea to life. It was something I envisioned to change the world lol. It obviously ended up being far from that but I’m pretty chuffed what I managed to achieve.

I am a bit upset I didn’t manage to get a distinction for this project since I felt like I had put my blood sweat and tears into it but I did obtain a merit at 64 points. It’s not too shabby but I went into my dissertation with the wrong mindset of “this must be groundbreaking and perfect” as per usual and I spent far too long looking for and changing my idea every single time I found a bit of paper that opposed what I wanted to do. I wasted so much time on getting stuck reading hundreds of academic reports trying to understand where I could find something “unique”. The whole idea of this research project I’ve now learn is NOT to find anything groundbreaking it’s just to pick something existing, change a variable, something small or find a weak point or area previous research has missed and do your diligence in building upon what is existing. You do this through the whole research process and what York mark you on is your progression from your academic research, literature review, to your methodology and conducting your experiments in a way they have explained to you exactly how to do all the whilst analysing and critiquing the new information you are obtaining each step of the way.

It does NOT matter what your project is, how unique it is, how complex it is or how good your results are. Your code and results can be shite for all they care but if you conduct the research exactly how a researcher would which is to base all your decisions on previous research, to base your evaluations on all your produced outputs and see everything through a critical lens at the same time accept your shortcomings you will do well! Coming from someone who is from a creative field – THIS WAS SO HARD. I just wanted to discover some mega original groundbreaking concept and then build it to life with the minimal skills I had lol.

So initially, the idea was to recommended inspiring stories of women who were in STEM careers to improve women’s perception of STEM occupations. However, because I spent 2 months and couldn’t get my ethics data collection approved I had to use a public dataset that was restricted and not include everything I needed in order to move forward. If I had resubmitted my ethics form and got it rejected again it would be 10 weeks out of 16 weeks just gone. Leaving me 6 weeks to do everything!

πŸ’‘ Tip
I recommend choosing a subject you’re well versed in using tools you already know. It’s normal to need to learn a few new things here and there but not an entire research area. Don’t be stupid like me and pick something you’ve never done! Although I loved my project towards the end of it and I highly regretted choosing something to learn from scratch I wouldn’t change what I did as I learnt so much but it made the entire few months very difficult to progress as quick as I wanted.

Ethics Form

In order to collect personal data you need to fill out an ethics form and send it to the ethics community at uni to get it signed off before you’re allowed to collect any data.

So this bitch F’d me up and was a complete disaster. I spent so long thinking my supervisor had sent off my form because I had sent it to him in the RRP module prior. I was waiting for a response for weeks then when I asked the supervisor about it and he said I had to send it myself?!?? I think other people’s supervisors sent theirs off, so it’s confusing as everyone is different. I sent it once about 6 weeks in and it got rejected. Then I sent it again and then at the 2-month mark, it got rejected again. FML. I was already halfway through my project without collecting any data I couldn’t start anything and was starting to panic big time.

πŸ’‘ Tip
I highly recommend being on top of your ethics form. Jump on your supervisor, chase them, and ensure yours has been filled out and passed on correctly – AND DO IT EARLY! You also need to be extremely specific with why you require the data, keep checking in with your supervisor. An even better thing I would suggest is to opt for an existing public dataset. There are barely any ethical concerns about these so you don’t need to wait for approval you can start A.S.A.P. The wait, rejection etc. can really delay your project. For public datasets I initially looked on kaggle and all the public places google and uni shared with us but I found my dataset via ChatGPT. I told Professor ChatGPT all the requirements of what I wanted in a dataset and boom linked me to two in 2 seconds.

The Supervisor

You will get allocated a supervisor who is aligned with their experience and your project …or whoever is available. You don’t get to pick yourself. Your supervisor is there to guide you through the dissertation not help you with the actual content etc. but to make sure you do your best, are on task and pass.

You will be advised to contact them in the first week to book a meeting. Depending on who you get you might have someone who needs a bit of prodding or you might get someone who was on the ball. Mine was kind of in-between, kind of needed prodding but was on the ball and present when needed. Some people got a very unresponsive Supervisor and complained and were able to change theirs – so don’t be scared to do this too!

I told my Supervisor that I needed motivation help by him panicking me. Which is did through out xD

My Project

So I’m not completely sure what the other modules require but since I was on the AI path I had to produce an AI-driven project, whether it was heuristics, neural networks, inference…idk what’s out there it could be anything.

I wanted to build a recommender system that recommended Role Models with the intention of focusing it on Women in STEM to attempt to tackle the problem of inequality in STEM occupations – or well use the idea πŸ’‘ as a basis. I was technically level 0 knowledge in recommender systems. I just knew what they did πŸ˜…. In the end I found a lot of papers on Career Recommender Systems (CRS) and based my model on a simpler one. I found that many research papers on CRS did not highlight gender disparities which would be more apparent in some career and fields like STEM. Thus recommendations could be imbalanced if gender was taken into consideration. But these could also be imbalance even if gender wasn’t since there may be other attributes that link gender beyond gender itself (I mean now there’s a few gender types you can align with this area is super blurred). So, I decided to focus on traditional genders, for now, just to understand the biases in the model.

I built it using manual methods of Python Sci-kit Learn Nearest Neighbours (NN) and not any ML tool since I wanted to manipulate it and make sure I knew what WTF was going on in the code lol. Then I used this to build a User-Profile based Recommender System that basically showed even though the input dataset was balanced in genders due to some careers being more male dominated the recommended Role Models in the STEM fields were all biased towards more male. So input -> output results. So I built a way to equalise the NN algorithm where it split gender and recommended neighbours in each split and then re-joined these back up. This helped balance the genders. So I was able to compare these two models by measuring the variation of recommendations. I actually found that equalising genders did not change variation much but by not doing so the recommendations were highly biases. This showed it was important to focus on reducing imbalance in dataset especially for Career Recommender systems where there is a huge gender disparity in certain jobs.

πŸ’‘ Tip
Choose a project, tools, and area you are very comfortable with and not something brand spanking new in. It also helps if you’re super interested in your project because 4 months is a hell of a long time to be doing something repeatedly and reading a shit ton of papers on it! Although I was a complete n00b at my topic I was luckily super interested in it and found all my research very exciting and easy to digest.

Project Management

I am literally THE most unorganised person I know. I have 3 alarms for every single important event I need to attend. And, well, this entire project is hefty!! I would have died without a Project Management (PM) tool I’m telling you!

If you struggle with organisation and motivation and getting stuff done well in the IRP you need to be self motivated for 4 months. There are multiple different areas you need to complete from research, writing, programming, testing, writing up results, evaluating etc so this can pile up and can get overwhelming.

Something Uni get you to do is a Gant chart to help plan out your days and work required. What I found even more helpful was to use a project management tool one with tasks, subtasks, tags and deadlines.

If any of you use project management tools in the workplace or even if you don’t these are tools to break down your projects into smaller tasks and those tasks into even smaller subtasks which you can then schedule into a calendar. This honestly can help with your overwhelm SO MUCH. As once you start planning these tasks into specific days, weeks etc all you do is check your calendar and look “oh today I’m working on X, Y, Z”. It stops you thinking about the big picture, getting anxious and overwhelmed by the amount of work you still have to do…. well it helped me anyways. Also always give yourself double the time you think you will need *chuckles*

I used Clickup as my tool since it’s free and I’ve used it before. You can find my actual IRP Gantt chart here just scroll to the left to May 2023 time. Clickup actually automatically creates the Gantt chart for you depending on what deadline dates you put into each task and project πŸ™‚ I actually planned ahead for the entire 4 months but also when things ran over or I didn’t need enough time I would adjust the bars so It reflected in real time the capacity I was allocating to each part. This helped build out the future capacity volumes for the future stuff I hadn’t done yet. It was a live working timeline that I updated weekly. If you couldn’t access the live one here’s a screenshot snippet of it here:

Here is also a screenshot of my Kanban board which is what the Gantt chart automatically generates off of. A Kanban board is just like a board with a couple of columns which represent the status of your sub-tasks. All sub-tasks are in your TO DO list which is self explanatory. When you start doing them you drag it into the IN PROGRESS column and then eventually the DONE column. You click into each task to describe what needs doing you can even add images, attachments, comments, check lists and also link it to other tasks that it may depend on!

Here is where I added all my links to all the research papers or websites too and their reviews/comments. I found that it helped so much with my natural disorganisation by keeping everything all in one place, logging it all and putting times against everything.

There are many other ones I now have moved on from Clickup and currently use Notion which I’m OBSESSED with. Notion is better for note taking and a Kanban board but it also has these bigger projects like clickup but the interface is a lot more slick especially on mobile app.

πŸ’‘ Tip: Use a project management tool to manage your entire research project. There are so many that are free. You can try out: Trello, Asana, Clickup, Notion, Monday

How I Felt Post-Hand in

I had a huge sense of relief the week of hand in. Like I was euphoric AND SO EXCITED about the future and all the fun things I could do…. then after for about a month I felt like something in my life was missing. I felt empty almost like it didn’t happen. What I really wanted at the end of completion was to feel like I had achieved something big and I had felt it for a bit but then time passed and i just felt the same as I had always done – constant imposter syndrome.

It made me reflect the parts that I enjoyed about uni that was rewarding and oddly enough it was writing the blogs, having other potential and existing York students reach out for support and help. That made me feel like I was actually making a difference in someone elses life. Me pursuing a Masters and just building on myself just wasn’t as satisfying as i had hoped.

I always felt like it was a sacrifice which was the hardest thing for me becsuse i was sacrificing my social and personal time with my partner, friends and all my energy for what? To gain more skills to do what?

What this taught me was it’s not about destination its about the progress. It was so gradual that I was always seeking this end goal but in the progress I slowly change towards becoming that person I always wanted to be which was someone that set goals, was consistent was organised and motivated and able achieve whatever I set my mind too whilst helping others in the process of my own development. It made me solidfy the things thay truly motivate me which was sharing my own experiences to instill confidence in others so they dont make the same mistakes I do. Be the person I wish was out there guiding me.

Anyways I was still sleeping at 2-3am for nearly a month after hand in since I was so used to the cramming of the late nights and I was often up feeling “on edge”. It had messed with my mental health , sleep and mood. I had gotten used to being cronically stressed it just felt normal to be like this. I am not condoning these feelings I am just sharing how I felt throughout the entire journey. After a month or so of feeling a bit blue I have since gotten back into all the things I loved doing like my hobbies like sewing, painting, playing the guitar. I have once again found joy in life in the little things that make me happy.

It’s now about 4 months after I’ve handed in and I still can not believe I’ve finished my Masters degree! It was all but a dream about 3 years ago and now life continues as it once did. I have learnt so much from this course not just computer science but re-learning how to learn, how to manage my time effectively and balance life. It’s helped me build so much perseverance and resilience and given me so much more than just a degree. Heck, I even found out I was dyslexic because of it! I feel a lot stronger because I feel like I know myself better and now I know I am able to set big goals and work towards achieving it – even if it takes a few years of elbow grease. The completion has boosted my self esteem in what I am able to achieve if I put my mind to it πŸ™‚

I’m planning on building those How to Build a Website Workshops I keep talking about in all my posts and never doing (heh heh). I will keep you all updated!

My partner and I went on a once in a life time holiday to Costa Rica and it was the first holiday in years where there wasn’t a deadline looming and it felt so FREE! We did so much hiking, saw so many animals, met lots of people did so many different activities.

This is me in a waterfall in Manuel Antonio, Costa Rica… hehe

Me on a 1600m Zipline over the Cloud Forrest, Monteverde

I want to say good luck to all the rest of you no matter where you are at! Whatever your goals are for doing this course there is light at the end of the tunnel πŸ™‚ You will have no doubt learnt so many new things about CS or even yourself and I just want to say thank you all so much for coming on this journey with me. I believe in you. Keep going, you’re so close now <3


If you found this post useful please support House Ninety Two: Buy Me a Coffee at ko-fi.com