Productivity & Self-Study

I first discovered the Goldilocks Rule in the book Atomic Habits by James Clear where he says:

“The Goldilocks Rule states that humans experience peak motivation when working on tasks that are right on the edge of their current abilities. Not too hard. Not too easy. Just right.”

James Clear

When you take on projects or learn new concepts that are ‘just right’ for your current abilities, you can more easily enter the ‘flow state’ while working where you’re so focussed on what you’re working on that you lose track of time and don’t get distracted by any of the usual things that pull at your attention (ahem, Instagram).

Similarly, Scott Young proposes the 85% rule – a way of determining how you should adjust your learning based on your success rate. He says that, on the one hand, if you get every exercise or task correct while learning, you’ll think you don’t need to make any improvements and you’ll probably overestimate your skill level in the topic. On the other hand, if you’re regularly getting exercises or tasks wrong then you won’t learn enough to be able to improve and your motivation will fizzle out.

The 85% rule is very useful for tasks that have an objective measure of success – such as math problems or coding exercises. In those cases it is clear when an answer is right or wrong and you can ensure you take on exercises where you get 85% of the problems or exercises correct.

How to Apply the Goldilocks Rule to Data Science Learning

Data science is a huge field and there is a lot to learn, with new developments being made and published daily. While on your data science learning journey you may find many new courses and tutorials on topics you know and many that you’ve never encountered before. It can feel much more comfortable to stick with beginner content until you feel ‘ready’ to take on more challenging concepts. However, if you do what’s easy or what’s comfortable, you won’t be able to improve and figure out where your weak areas are.

I encounter this a lot when taking a new course or starting a new book. Many courses and books will often try to cover a broad range of topics that go from beginner to intermediate and sometimes even to an advanced level. The mistake I make is that I think a course must be taken from lesson 1 or books must be started from chapter 1. This is definitely not the best approach. It is difficult to maintain motivation when everything feels so easy. You’re not being challenged or pushed to improve yourself and so you will most likely not follow through with the book or course.

On the other hand, if you take on an advanced NLP course when you haven’t even covered some of the basics of text analysis or neural networks then you’ll probably feel very demotivated by how difficult the course is and quit.

The better approach is to identify what your current skill level is and take on new concepts, exercises or projects that are just at the boundary of your capabilities. Not too hard. Not too easy. Just right. Resist the urge to remain in your comfort zone where things are easy or that you know very well, or to take on things that are too challenging and you have no foundations around the topics that are covered.

How to Use the Goldilocks Rule to Stay Motivated

Something I am always trying to figure out is how I can stay motivated to do things over a long period of time. Even though I love data science and I love learning new things, I still feel my motivation levels running low at times.

While I have not found something that instantly makes me motivated all the time, you can use the Goldilocks Rule as one tool in your toolkit to help you maintain motivation over long periods of time.

The reason that this works is because as human beings we are hard-wired to avoid pain (i.e. very challenging cognitive tasks in this case) and we absolutely hate being bored (which is why so many of us are addicted to social media – dopamine, please).

By working on tasks that are at the edge of our capabilities, we get constant feedback on how we are doing. This means we get a dopamine boost for the things that we can actually get right but are still challenged enough so that we don’t get bored or frustrated.

However, an important part of the Goldilocks Rule is keeping track of how you are doing and measuring you progress. It’s easy to subjectively look at everything you’ve learnt over the course of a day/week/month and say “everything was terrible” (I don’t know about you but I tend to rate myself very negatively). However, you may find that if you looked at your progress over time, you may actually be improving quite a bit.

The best way I have found for measuring your progress is using a scoreboard that I discovered from Tina Huang. She prepares a spreadsheet with all the days she plans to study and she will mark whether she was able to get the things done that she said she was going to do. In this way, she is able to clearly visualise in a chart how she has been progressing over time. This gives immediate, clear feedback on your progress and is essential to staying motivated in the long-term.

I have a definite love-hate relationship with productivity. To me, being productive means that I am actively working towards my goals and dreams. However, productivity is not something I can objectively obtain every day. Instead, it ebbs and flows based on so many factors such as diet, sleep, sickness, weather, connection, conflict, my own internal narrative, and the list goes on.

It is almost impossible to control exactly how productive you will be in a given day, week, month. Because of this, I use a few hacks that give me the best possible chance to have a successful day. I’m not perfect – I don’t cross-off all my to-do’s every day, I procrastinate (all the time) and I find it incredibly difficult to reach ‘flow’ state.

The secret is to not try to reach perfection every day but instead to take just 1 step forward, and then another, and then another.

Setting up my Environment

There are 3 aspects to my environment: my physical space, my digital space, and my personal body.

Physical

I try to make sure my physical space is tidy, clean and inspiring.

There are a few plants on my desk that I love looking at, and they quite literally bring the space to life. Studies have actually proven that house plants improve concentration and productivity by 15 percent! I don’t know about you but I’ll take that 15 percent, thank you.

I have a good chair, a desk that allows my arms to rest at a 90 degree angle and a monitor at eye-level so I tick all the ergonomic boxes there.

Digital

Before I log off for the day, I try to make sure that all distracting apps and browser tabs are closed. There is nothing that ruins my morning more that sitting down in front of my computer, turning on my monitor and the first thing I see is Twitter, Reddit, and 20 other open tabs. The first thing I’d want to do is some time-wasting scrolling and then… ugh, can I just go back to bed?

Thats why I close all open tabs and save any interesting articles or links to my Notion at the end of the day. If I really need it, I can always go back to it again. Scrolling through my browser history is a big time kill.

Personal

This is for making sure I personally feel good before I sit down to work. Remember that this will be different for everyone and this is what works for me.

It is non-negotiable for me to get dressed every morning. I don’t wear anything formal, but it needs to be different to what I was wearing when I woke up.

My mental and physical health have been sore topics in the past. I have often found it difficult to imagine and accept that maintaining a regular routine, eating well and working out can actually make any significant difference to my life.

When working from home I have found it too easy to let these things slip. I went through a phase where I would only wake up at the very last minute before it was absolutely necessary for me to be sitting at my desk, I didn’t workout at all and I ate way too much take out. Most of this came from a place of plain laziness and, needless to say, it did not do anything to help my mental or physical health. I frequently felt unmotivated and sluggish during the day.

I have realised over time how important it is to prioritise these things over all else. Now, I try to workout 4 times a week and eat a reasonably healthy diet. I have to regularly remind myself that working out, taking breaks, and resting when I need it are ALL productive activities.

Time Blocking in Google Calendar

I was first introduced to the idea of time blocking by Cal Newport in his book Deep Work. Cal recommends using pen and paper for this but I find that my perfectionism (e.g. needing all time blocks to be an equal size and symmetrically spaced) and the need for everything to look ‘neat’ holds me back so I use Google Calendar instead.

At the beginning of the week (preferably on a Sunday), I go into my Google Calendar and review my commitments for the week, making sure everything is actually in my calendar. Then I fill in my meal times and any other activities like exercise and leisure time. I also leave some blocks blank and leave space between meetings so that there is still some flexibility in my day.

This allows me to realistically see what blocks of time I have available for deep work and I will fill those in. I try to organise my day so that I can set aside at least 3 hours for deep work. This isn’t possible every day but a few 3-hour-chunks per week is achievable for me. I do this because I have come to realise that content-switching and multi-tasking is almost impossible for me. I cannot get any amount of quality work done in short 30 minute slots here and there.

The combination of time blocking and deep work really helps me to avoid procrastinating because I always know what I should be working on, when I should be working on it and how much time I have set aside for it.

A good thing to remember is that no plan is set in stone. If you don’t feel like doing something that you set in your calendar, move it! If you wake up with a raging headache and need rest then listen to your body. Your future productivity will thank you for taking the rest!

Each morning, before I start my work day I review my calendar and write out the tasks I will complete according to the amount of time I have set aside for the day. I try to break the tasks down into just 1 action. Instead of having a task for ‘write blog post’, I’ll have several tasks like ‘write outline’, ‘brain dump in each heading’, ‘brainstorm headlines’, ‘collect 5 research links/resources’, ‘write first draft’, ‘edit first draft’, ‘design blog banner graphic’, ‘publish blog post’.

Each task is achievable and there is no ambiquity around exactly what I’ll be doing. I also ask myself which tasks are essential for me to feel that I have won the day. If things don’t go as planned and something else comes up at the last minute and if I at least do these essential tasks then I will feel like the day is still won.

The Pomodoro Method

This is something I use every day on the Forest app. Once I am ready to start working on a task, I will set a timer on the Forest app. If I am feeling a lot of resistance to the task then I will set a short 25 minute timer to just get me started with it and usually I don’t mind setting another 25 minutes and another and so it goes. Sometimes I just do a single 25 minute pomodoro and that’s it. If I still struggle to focus on the task, I move on to something else.

If I don’t have much resistance then I will usually default to a 50 minute timer with a 10 minute break. While I’m doing some tasks, my mind often wonders to other things like “I want to go for a walk, it’s such a beautiful day” or “I should make coffee, I’d feel more productive with a cup of coffee” or “I’m suddenly feeling hungry, I should get a snack”.

The pomodoro technique is so powerful for my wandering mind because I will look at the timer and think “NO, only 20 minutes to go. I just have to stick it out till the end and then I can go for the walk/coffee/snack but I just need to do this for 20 more minutes”.

During deep work, I still use the pomodoro method but I the breaks are shorter and usually only involves a quick stretch and maybe a trip to the kitchen for more coffee water. I try not to do anything that will significantly distract me or turn my attention in some other direction because that defeats the purpose of the deep work session.

I also sometimes set timers for non-productive tasks. For example, if I want to go on social media or check emails I will set a short timer for it. This makes sure that I don’t fall into an endless pit of scrolling through social media or replying to emails for hours because there is a time cap and then it’s back to work.

Co-working Communities

During the pandemic, like so many others, I wanted to feel part of a community where I could get back the connection I felt I had lost during this time. It took some time to get used to it and gain the confidence to put myself out there in a new community and make connections with people. But I am happy I spent the time on it because I have found some amazing communities of people with similar interests to me.

A big aspect of this is that these communities can then form part of your accountability system. Talking about my goals and putting them out there takes them out of my head and forces me to take action to achieve them and I know that I will be supported and held accountable to them too!

Lastly, there is the virtual co-working. You could do this on a call (with or without video) using a platform like zoom or you could even join a live stream on Twitch under the ‘Co-Working’ tag and work alongside the streamer. Twitch co-working streams has started to get very popular in the last year or so.

After trying out various forms of co-working I find that video calls are the most effective. There is an element of pressure to work because the other people on the call can see you and you don’t want to look like you’re procrastinating. Besides, they’re also working so you’re not alone.

I hope you found this post helpful! Let me know in the comments below what the number 1 way is that you stay productive while working from home.

Reading a textbook for self-study is very different to school or university. Back when I was in school, I read the textbook for the sole purpose of doing well on a test and it didn’t really matter if I couldn’t apply the concepts I was learning about. However, I have come to realise that this is the most critical part of self-study.

Over the years, I have developed a few helpful and effective strategies for reading technical textbooks and I want to share them with you. Also, if you didn’t know – I have created a 6 month data science curriculum and I will be applying these strategies as I read through a few of the technical textbooks in my curriculum.

Annotate & mark it up

If you are able to annotate or mark up your text book then you definitely should. Annotating a text book almost forces you to read more actively. You approach the text with the purpose of identifying what’s important and what’s unclear and it prevents you from just mindlessly reading words that mean nothing.

I prefer to just use a pencil when annotating a book, but that’s my personal preference. If you like using highlighters then go ahead! I annotate like this:

• Underline or place an asterix next to important concepts & paragraphs – most of the time this is subjective and is just what I think is important as I read
• Place a question mark next to things I don’t fully understand with a little note on what question I had in my mind so I know to revisit it or spend more time on it later
• If something is very complex I will usually scribble some notes or try to break down the concept on a small piece of paper that I insert into the book at the position I am reading. This helps to identify what I DO understand and when I revisit the section to make further notes, I can focus my attention on the parts I DON’T understand. The worst thing you can do is say “I don’t understand section 5 at all” because that doesn’t help you. What, specifically, about section 5 didn’t you understand? Often, all you’re missing is a small piece of the puzzle that makes everything click into place.
• Lastly, if I come across anything that I know I want to come back to for a project or an implementation, I’ll add a little sticky flag on the edge of the page so that it sticks out to me when the book is closed.

Don’t take notes on everything

This is probably the hardest thing to remember. Throughout my time at school and university I was always a compulsive note-taker. I took extremely comprehensive notes on absolutely everything to the point that my notes became more of a hindrance than a help. I learnt the hard way how to take better notes.

When self-studying a textbook, this is even more important to remember because you are probably not going to take a test on the material. Instead, you want to approach the text with the goal of becoming a practitioner (more on implementation later).

On the topic of digital vs handwritten notes, I prefer taking digital notes. I keep all my notes in an app called Notion and it allows me to capture almost anything such as links, embeds, formulas, pictures, code snippets, etc.

These are some of the benefits I have experienced when taking digital notes:

• I can type faster than I can write
• I can jump around topics – my notes do not need to be sequential or perfect the first time I write them
• I can fill in or add to my notes, and easily delete irrelevant or badly written notes
• It does not come with the side effect of hand cramps after an hour of writing 😬

As I write my notes, I try to keep these points in mind:

• Take notes on the higher level material – don’t get stuck taking notes on every detail you read. That’s what the book is there for, don’t rewrite it.
• Ask questions in your notes (and find answers for them in other texts, online articles, video’s, etc. and link to them in your notes)
• Link concepts together in your notes from other parts of the book or from other resources (similar to the question/answer method above). I believe that true understanding and learning comes when you’re able to link together the things you are learning
• For complex concepts that I struggle to understand, I would get out a blank sheet of paper (I go analog for this – pen and paper are the best for complex and challenging material) and pretend I’m explaining the topic to someone, starting with what I do know and then filling in the gaps in my knowledge with the material in the textbook (see the Feynman Technique).

Implement fast

I am a big believer in project-based learning, especially for technical topics. There is an excellent quote by Stephen R. Covey in his book The 7 Habits of Highly Effective People:

To learn and not to do is really not to learn; to know and not to do is really not to know

Stephen R. Covey

When going through a textbook I am constantly asking myself how I can apply what I am learning. Before I have even completed a chapter, I will start looking for data, python libraries, and example code that I can use in a project that applies the algorithm, model, system or high-level concept I am learning.

The faster I can get to the project-building phase, the better. Once I start implementing, I usually end up with a list of questions that did not even occur to me while I was reading the textbook and I quickly realise which of the concepts I didn’t actually understand and I’ll go in search of answers. Sometimes the textbook doesn’t explain something in a way that I understand and reading it somewhere else, explained in a different way, can suddenly make things click.

Iterate & expand

Think of your study as a iterative process. You may not cover every detail or completely understand everything you read the first time you read it. However, you should understand just enough to get to apply the concepts in practical projects, as mentioned in the point above. Once you start working with the material in a practical way you’ll be able to identify the gaps in your knowledge and then you can refer back to the book to expand on those areas and fill in some of the gaps.

During each iteration you should also return to your notes. Add new ideas and remove points that don’t make sense or aren’t relevant to keep notes on.

Remember not to get too attached to your textbook. There are hundreds of resources out there for learning something and expanding your knowledge requires different perspectives or different ways of teaching something. Explore.

I hope you found these strategies helpful. Leave a comment below or connect with me on Twitter and let me know your thoughts on self-study and reading technical textbooks, I’d love to hear from you!

Welcome to the next 6 months of my data science learning journey!

After being heavily inspired by Daniel Bourke’s AI masters curriculum, I have created my own learning curriculum that I will be following for the next 6 months.

Will this be a challenge for me? Yes.

Will the content be difficult and hard to understand? Probably.

Will I hit roadblocks and hurdles along the way? Almost definitely.

Is it achievable? Absolutely!

Is it for everyone? Maybe. But you get to decide what’s possible for you. If your dream is to learn data science then accept the challenges that await you and get started!

Please remember that everything here is my own opinion and my own way of doing things. You may not agree with it and that’s totally ok! You can create a learning curriculum that aligns to your own goals and beliefs.

Why am I doing this?

Learning data science and machine learning is no small feat. The field is advancing rapidly and it takes continuous study just to keep up. My goal with this learning curriculum is to build up my foundational knowledge and become proficient in machine learning and deep learning. I also want to be able to take on more challenging projects or pick up a new concept faster in the future.

This is not intended to be a beginners guide or something that you should do if you are just getting into data science or programming for the first time. I have a masters degree in statistics so I am already quite comfortable with the math and stats that form the foundations of machine learning. I have also been coding in python for a few years so I’ve already got a good understanding of the basics of programming in python.

However, I have quite a few gaps in my knowledge when it comes to applying machine learning algorithms to real data in python and I also have not properly explored deep learning so that will all be completely new to me.

I am not trying to get through the content in these courses and books as fast as possible just for the sake of saying I did the course. So I will probably be skipping some sections if they don’t feel relevant or if the content is a duplication of something I already learnt in another course.

It is still very important to me that I understand the content and am able to apply it myself. Therefore, my curriculum has a very strong focus on projects. Most of what I learn will be applied in a practical (and realistic) way with new, real-world data sets (not the data given in the course/book).

I will also probably not be trying to get any special certificates as part of the courses that I complete. I don’t particularly care for the testing methods of these courses which are very far removed from real life. Thats why I’ll be completing projects on my own that will give me an indication of my competency and level of understanding as I progress through the content.

How much time will I spending on this?

I roughly estimated how long I think each course will take (including all the projects) and would also like to spend about 4 hours per day, 6 days a week studying. From this I felt that 6 months would be a challenging yet achievable amount of time to complete this curriculum.

Why did I choose these courses/topics?

Once again, I got most of these resources from Daniel Bourke’s website. He has a comprehensive list of some of the best resources to learn data science, machine learning and deep learning. I encourage you to check it out if you’re trying to figure out what kinds of resources are out there for learning this stuff.

Also, the courses I chose can all be taken for free, except for the Udemy course which I got at a ridiculously discounted price. The only caveat is that the Cousera courses do not provide a certificate if taken in their free ‘audit’ mode and there may be some limitations. The only other thing I paid for was the Hands on Machine Learning Kindle book and the other 2 books were given to me as a gift.

For my own curriculum I created 2 sections which I then allocated the material to:

• Foundations & machine learning – covering the general approach to machine learning problems and the various algorithms used in machine learning (such as k-means clustering, regression, decision trees, support vector machines, etc.)
• Deep learning – covering topics related to neural networks, NLP, computer vision, hyperparameter tuning, optimisation, best practices, etc.

How am I tracking my progress?

I created my own Notion page with a kanban board and a calendar. I estimated the amount of time I need to get through each resource and gave each project a deadline. This will be my guide as I progress through the curriculum.

My aim here is to get rid of this question that I often ask myself when learning new things: “ok, what should I do right now and then what’s next?”. Using this curriculum and guide in Notion means that I should always know what I need to be working on at any point and what will be coming up next. A little structure makes me feel better 😊

If things change during this time, then I will simply adjust my timelines in Notion. It’s important to still be flexible to life events, disruptions, other interests and the possibility that some concepts will just take longer for me to grasp than others and I won’t know what I might struggle with until I start.

Each project that I complete will be added to my GitHub and made into a blog post. Project-based blog posts will go up at least every other week. I’ll also be writing other posts on some of the key concepts I’m learning so that I can reinforce the knowledge by distilling the information down into (hopefully) easy-to-read blog posts.

I will also be posting my experiences, interesting facts, new projects and challenges I face to Twitter everyday. There will also be a weekly check-in tweet where I will give an update on my progress in the curriculum and what I’ll be covering during the week.

The Curriculum

Alright, that’s enough chit-chat. Here are links to all the resources I’m using in my curriculum:

Foundations & Machine Learning:

• Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems (Book)
• Python for Data Science and Machine Learning Bootcamp (Udemy course)

Deep Learning:

• Practical Deep Learning for Coders (Fast AI Online Course)
• Deep Learning Specialisation (set of 5 courses on Cousera)
• Full stack deep learning
• Hugging Face course (NLP)

Additional Resources:

• Machine Learning by Tom M. Mitchell (Book)
• Deep Learning by Ian Goodfellow et al (Book)

Here is a little snippet of my study plan in Notion with timelines:

I hope that by sharing my curriculum you will be inspired to also make a commitment to learning and building in data science!

Follow me on Twitter and join me on my journey. Send me a DM or comment below and let me know your thoughts, I’d love to hear from you!