Predicting my own finish time with less than 45 seconds error.

Image by Marvin Ronsdrof, Unsplash

It has always bothered me that finish time prediction in running race is so basic and inaccurate. It disregards factors like the elevation curve of the race and the fact that most runners get tired as they get further into the race. It’s also unable to estimate your finish time before you start the race.

As a spectator you can follow a list of runners and see how they progress throughout the race. This is possible due to checkpoints along the track that runners pass every 5 kilometers. This…

Machine learning for running — Machine learning & results.

Photo by Candra Winata on Unsplash

This is part 3 in the educational finish time prediction series:

In this post, I will show you how I managed to predict my own finish time (1:40:34) with less than 45 seconds error before the race even started. We will use some of our findings from the first parts to:

  • Implement features for our model
  • Train 5 different models to predict finish times at each checkpoint in the race
  • Evaluate our results

Why we need machine learning

In part 2, we concluded that average…

Machine learning for running. Feature engineering & Hubris analysis.

Photo by Mārtiņš Zemlickis on Unsplash

This is part 2 in the finish time prediction series:

In this post we will dive into pace analysis to see if we can understand the runners even further.

Let’s pick up where we left off:

Pace change — finish time versus age

Machine learning for running — Data analysis & Intro.

Image by Marvin Ronsdrof, Unsplash

Göteborgsvarvet is the largest half marathon in the world. In May, 2019 the race celebrated its 40 year anniversary, attracting over 50 000 runners and 200 000 spectators.

As a spectator you can follow a list of runners and see how they progress throughout the race. This is possible due to checkpoints along the track that runners pass every 5 kilometers. This is a fun way to keep track of your friends when you can’t see them. You will also see a projected finish time for your runners, which is based on their average pace so far.

It has always…

Solving a common real every day life decision making problem using KMeans clustering.

Image from Pexels

At this year’s Codecation, we had an activity planned to break up the programming sessions. As a group of 17 people, it’s hard to accommodate everybody’s wishes so we organized a session where we voted for what we wanted to do. The poll included 8 alternatives:

🚴‍ Biking
🏖 Beach
🧗‍ Gibraltar Rock
👟 Hike (mountains)
👀 Village watching
🕵️‍ ️Escape Room
🏸 Paddle
🍷 Wine Tasting

Everybody was asked to prioritize the activities between 1 (most preferred) and 8 (least preferred). This is a powerful voting system since you can also specify what you do not want to do…

A guide to annotation, dependency parse trees and linking

In this post, we will be focusing on text mining and review analysis. We’ll roll our own deep learning implementation for relation extraction. Imagine you have a large pool of reviews about your company or your products and you want a quick overview of what descriptors or adjectives are mentioned about certain parts or features of your company or products. The end goal looks something like this:

Aggregated result for hypothetical headphone reviews

Performing this over thousands of reviews and aggregating this together builds a pretty powerful summarization tool that can be used to get a quick and thorough picture of what is said about a…

How many kiteable days are there per year at different locations?

A friday after work session in June, Landskrona (southern Sweden), Image by Author

Co-written by fellow ML-engineer & kitesurfing enthusiast Michal Stypa.

While thinking about the next kiteboarding session, we asked ourselves if there is anything we as data scientists can do to get the most out of our beloved sport? We were looking for tangible insights, something that would give us an advantage in planning when and where to head out for the next session.

For those not quite familiar with the sport, kiteboarding requires fair amounts of steady, on-shore wind. It is a simple but crucial condition. No wind, no kite. …

Backtick @ Web Summit 2018 — Developer impressions and insights

Web Summit is the largest tech conference in the world, attracting around 70 000 attendees from 159 countries. More than 1800 exciting startups and 1200 speakers guarantee you will never find yourself without something to do!

I often found myself wanting to go to two different talks or workshops at the same time, so planning and prioritization is key to survive. The venue is huge and if you think that your 10 minutes gap between two talks is enough for a coffee break and relocation to the other stage —…

Afternoon sesh, photo by Author

Codecation is a motivating hackathon and a small-scale conference focused on exploring and sharing exciting ideas with like-minded engineers.

This year, we decided to organize Codecation in Barcelona. Picking a location is always tricky. You want it central enough so that there’s good enough WiFi to support the entire entourage with high-speed internet without limitations, yet remote enough to provide that feeling of isolation required to achieve optimal team-building and productivity. We found a villa in Badalona which had everything we needed.

Formalizing Backtick Technologies AB

We have some very exciting news to share!

Late september, the formalization of Backtick Technologies as a company completed. Backtick Technologies will focus on creating exciting, impactful software in areas related to data engineering, machine learning, artificial intelligence and full stack development.

Oskar Handmark

All things data engineering & ML, Founder of and, co-author of

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store