Skip to main content

6 Ideas To Inspire Your Next Technical Conference Presentation

Giving conference talks is a rewarding activity that data scientists and developers can – and should – regularly take part in. Having given four talks during the recent Europython event, our team enthusiastically agree. Giving back to our community in this way leaves you with a warm fuzzy feeling. But what do you do when we are having trouble getting started?

Here are 6 suggestions to help inspire, to encourage those who are hesitant to start, and help make our tech community better with whatever they have to say.

Ondrej Urban - Trayport

Ondrej Urban

To start, let’s look at why you might be talking yourself out of submitting a proposal for an upcoming conference. This is common and – very importantly – not based on reality.

  • “I have only started in this field, barely know anything – who could possibly be interested in what I have to say?” Rest assured it’s not possible to make a talk too basic. There are always those yet to learn what you know, and you are not the only (self-proclaimed) beginner at a conference. If in doubt, keep XKCD’s Ten Thousand in mind!
  • “Everyone will see right through my lack of in-depth knowledge and I’ll crash and burn during the Q&A.” This is a clear example of Impostor syndrome and if you have encountered it, try not to listen. Keep in mind people come to conferences to learn, not to humiliate. Even better, some version of “Be nice to others!” has explicitly made it into the code of conduct of most conferences. Rest assured everyone is happy, supportive and will do their best to make you feel welcome and appreciated. In the worst case, the chair of your session will always have your side.

Now, let’s assume you’ve curbed your inner doubts and are keen to actively participate. What in the world should you talk about? Here are a few broad avenues to guide your thinking when choosing what to talk about. Let’s get inspired!

1. A New Tool You/Your Team Have Developed

Let’s get this out of the way quickly. For many beginners, this may sound like the only “proper” kind of a talk to give at a developer/data science conference but since they are rarely in possession of such a tool, it seems like they are out of luck. As we will see below this is far from the truth. However, if you or your team happen to have developed a cool new tool or a library, go out there and show it off!

2. A Problem Solving Use Case

Sometimes it might be easy to forget that software tools and libraries are primarily there for problem-solving. Can you think of an interesting use case for which you can use some of the broad range of tools available today? Do the research and tell everyone what you have found! 

For example, a few years ago, my colleague and I gave a presentation about finding connections between bad air quality and the prevalence of Alzheimer’s. We used the current versions of various software libraries for data wrangling and plotting data on maps. We found great benefit in both learning these tools and displaying the tools and the results to our audience.

3. A Cool Tool For A Particular Problem

This type of talk tries to answer the question “What is the best way to do X?” 

It is somewhat related to #2, but the focus is on the tooling. The use case thus tends to be straightforward, reducing the time needed to explain it to the audience. You can compare how different tools perform a task or focus on a single tool – the best one. In either case, don’t worry about this not being a tool written by you – give kudos to the authors and spread the word about their good work! 

Two of our Europython talks were of this type – in one we described the functionality of linear optimization library pyomo by finding the optimal seating of the guests at a fictional wedding. In the other, we demonstrated how to apply the automated machine learning libraries Keras Tuner and AutoKeras to a well-known problem of the MNIST digit classification.

4. A Loosely Connected List of Tips

During your career, you’ll spend a lot of time digging into code. You might even discover the inverse correlation between time spent debugging and the intricacy of the ultimate solution. As months and years go by, you’ll accumulate a set of tricks you’ll keep using – tricks to make your code run faster, code practices that make your code more readable, the structure you apply to projects you are starting, or anything else. You know who would benefit from your little tricks? Everyone else! Find a topic that connects them – however loosely – and make the lives of others simpler. An added bonus? If whatever annoys you does not have a solution yet, you can create your own tool and present that instead. No need to make it complicated. Have a look at Calm Code, where a big fraction of libraries presented were inspired by exactly this kind of little annoyances their authors were repeatedly running into (learn about them at a recorded pyData talk

5. Use Something You’ve Already Done

Have you presented something in the past? Think in the broadest sense possible – presenting the findings from your last JIRA research ticket to your colleagues, talking at an internal company event, a meetup. Whatever the case might be, you now have a great resource at your disposal – a tried and tested story. You know what worked and what didn’t. You might have ideas or received feedback on how to improve it or what to add. And you have done at least one practice run, namely the original presentation. 

Submitting a talk proposal is a no-brainer. In addition, you can now fine-tune how much additional work is needed. A significant revamp to make it into something new and even more exciting? Straight recycling? Both extremes and anything in between is possible!

6. Lightning Talk

In our final example, we turn to an almost opposite problem – you have a talk in mind but your proposal was not successful. Not to worry, you can still get your message across as a lightning talk, a format that’s been very popular. A lightning talk session consists of short informal presentations of up to 5 minutes that follow each other. The sign-up lists for these sessions are usually only available during the conference and with a bit of luck, it is relatively easy to secure a spot. The timing tends to be strict, but that gives you a chance to be creative and minimalist in getting your message through in this short amount of time.

Giving regular talks at developers/data science conferences is something we all should do. Even though it might seem intimidating for a beginner, we have discussed how most of that is only due to impostor syndrome and how you should just go ahead and propose a topic. Given the broad scope of the tech field, there is an infinite number of topics you can propose – and if you’re struggling, try one of the six avenues above. Before you know it you’ll have your idea.

See you at the next conference!

Ondrej Urban
A former astrophysicist, Ondrej is a data scientist on Trayport's AI team trying to teach computers to help people with trading. He spends his free time reading books, cycling, or refreshing his weather app to plan his next ride.