How to Prepare your Next Talk
BlogConferencesLearning data scienceposted by Diego Arenas, ODSC September 7, 2017 Diego Arenas, ODSC
Editor’s Note: Interested in speaking at ODSC East 2018 in Boston? Apply Here
Here is some advice on how to best prepare for your next talk at a data science conference.
Even though you are an expert in your field, the delivery of your message is just as important as the message itself. It is not enough for you to transmit your information, you also need your audience to comprehend that message. Think of two speakers, one reads the slides and another makes jokes and tell anecdotes about the content. Which one do you think the audience will not only enjoy more, but better engage and learn from?
There are two things to consider when preparing your talk. First, the level of expertise of the audience regarding the subject you are presenting and second, the potential takeaways for the audience.
Try to figure out how much your audience knows about your topic, in order to define the scope and content of your talk. Include beginner and intermediate takeaways based on the expectations people may have attended to your talk.
If you don’t know how much the audience knows about your talk, then just ask them. After introducing yourself and the topics, ask them what you need to know about their familiarity with your topic, and proceed based on this information.
Think of your talk as a story, after all you are the storyteller.
- What is the message you want the audience remembers after your talk?
- Use that message to structure the content.
- What are the arguments to conclude that?
- Think about one key message founded upon 2-3 supporting ideas.
Divide your presentation into sections. Talk through the topics you are going to cover. A slide with the agenda is an useful guide to the presentation and also you can add a transitioning slide between sections. You can start with three sections: introduction, development, and conclusion. You can add more sections if you want such as background or demonstration.
A good strategy is to start with the conclusions. It will provide context with which to narrate your story and make your audience comfortable. Don’t make the audience over-think what you are trying to say, be explicit and direct.
If you have an important idea to deliver you can use this technique:
- Say what you are going to say,
- Say what you want/need to say, and then
- Say what you just said.
You should consider if your audience will be neutral / supportive / against you? Your story must be coherent, concise, and sound.
To deliver your message use as many senses of the audience. Vision, hearing, tacts, etc. Use emotions if you can.
If you want to include demos, some people (myself included) record them in advance and show them as videos. This comes with pros and cons.
Pros: The demo has less chance of failing during the presentation (except if the video doesn’t work in the host computer); this will allow you to explain concepts clearly using the video; also it will give your audience a visual aide if you decide to share your presentation.
Cons: You will have to prepare the videos in advance and that means repeating the recording until everything looks smooth. That eliminates the excitement of a live demo. Use big fonts for the video unless you know the resolution of the projector, there will be no opportunity to change it afterwards; bear in mind that your presentation will be 10x or more times heavy than a bunch of text slides.
Think of the top three ideas you want them to remember once they step out the room. And make sure you repeat them enough at different parts of your presentation. The ideas you want them to remember are different from the takeaways from your presentation. For example, you want them to remember that method A is better than method B when C happens but worst when D happen. The takeaway here is to be aware whether the C or D case is happening, not the methods.
Include real experiences in your talk whenever possible, your own or from people you know. Empirical experiences make the message more tangible and believable. “When we did this…”, “This didn’t work for us because…”, “I tested this using…”, etc.
Then decide your style of presenting. What helps you to feel confident? Having a lot of interaction with your audience? Telling stories, making demonstrations? I have been to several meetups, presentations, seminars, conferences, etc. and I have seen distinctions between the way people presents their talks. I can group them in three styles: in the clouds, the wise humble, and the slides reader.
Prepare more content than the talk allows. Sometimes you will have to complement your presentation if you have extra time. Include some slides after the final one with detailed explanations to questions that may arise at the end. Think about the questions you could receive and prepare slides to explain those you think are more important to explain.
You will not have the full attention of all the attendees, that is a fact. Attention decays with time. If you see your presentation boring the audience, make questions that will bring the audience to the moment forcing them to use their brains to reason about the questions you threw at them. “What you think will happen if…?”, “Raise your hands those who…” (this one is used more often than I would like to but it’s still worth it.)
Structure your talk into five periods. In the first, ease your audience into it; outline what you will cover and prepare them for the talk. Then use the next three, to explain your main findings and ideas. The focus and attention will go into crescendo during this time until you will lose them after approx 45-50 minutes, consider this time to make a break and use the time to summarize what you’ve covered so far. If your presentation is for 40 minutes, you will be fine, keep the first 10 minutes for warming up the brain and then transition to more complex matters in the next 30 minutes, you will have them at their peak of attention by the time you make your conclusions.
If your audience resembles a picture while you are presenting it means they are attentive and focused. If you see some movement on their seats, it might mean they are distracted or bored that will necessitate actions like asking questions, sharing a personal experience, or speaking louder.
Avoid using a monotone when presenting, emphasize your voice to highlight important parts of your talk. A lower tone of voice is usually well-more received, so use it for conclusions and important parts of the presentation.
Always be vigilant of the timer and be prepared with extra content in case you finish earlier than expected. You know your content, so don’t cram a semester’s worth of lectures into one presentation.
Practice your presentation before giving it. Present to someone with a similar profile as the attendees to your talk, and use their feedback to improve your presentation skills and the content of your talk.
Your presentation will have a structure and an action plan. So you can decide at any moment if you want to take a shortcut or to fixate on a single point. That way when you check the remaining time you can decide what to do, this gives you the freedom to decide where to go with the time you have left.
Some common mistakes to avoid before your presentation:
- If you will include live demos, time your presentation in advance.
- Make sure you have a wifi connection for your presentation.
- Be prepared to present from any device, so backup your presentation in different data formats just in case.
- Check the resolution and colours of the projector you will use.
- Make sure the fonts you are using are adequate.
- Make sure to include the references and credit sources.
- Maintain visual contact with your audience. It’s easy to get trapped between your screen and the projection of your presentation. You can receive this feedback if you practice your presentation in advance.
In summary, design your talk using the level of expertise of the audience, and plan some takeaways for the attendees.
Good luck with your next talk!

Diego Arenas, ODSC
I've worked in BI, DWH, and Data Mining. MSc in Data Science. Experience in multiple BI and Data Science tools always thinking how to solve information needs and add value to organisations from the data available. Experience with Business Objects, Pentaho, Informatica Power Center, SSAS, SSIS, SSRS, MS SQL Server from 2000 to 2017, and other DBMS, Tableau, Hadoop, Python, R, SQL. Predicting modelling. My interest are in Information Systems, Data Modeling, Predictive and Descriptive Analysis, Machine Learning, Data Visualization, Open Data. Specialties: Data modeling, data warehousing, data mining, performance management, business intelligence.
George R.R. Martin And Other Authors Sue OpenAI
AI and Data Science Newsposted by ODSC Team Sep 22, 2023
9 Open Source LLMs and Agents to Watch
Modelingposted by ODSC Team Sep 21, 2023
Conversational Data Analysis: Cutting Through the Noise to Find the Real Deal
Modelingposted by ODSC Community Sep 21, 2023