Getting Serious with Time Series
Time series—sequences of events or values that we can analyze to detect meaningful patterns—are a key element in data scientists’ toolkit. From environmental studies to marketing, they play a crucial role across numerous stats-rich fields, and empower practitioners to draw insights about the past and make informed predictions about the future.
This week, we’ve selected four enlightening contributions that cover time-series forecasting from different angles. Regardless of your experience level, you’re likely to learn something new.
- It’s important to get the basics right. If you’re taking your first steps in time-series analysis, Sameeha Afrulbasha’s introduction to the topic is a great place to start. It offers clear definitions of fundamental concepts, focuses on ARIMA (autoregressive integrated moving average) models, and explains how the different components work together.
- Bringing time series to life in Python. If you’re ready to roll up your sleeves and work with some real data, Leonie Monigatti makes the transition from theory to practice smooth and painless. This helpful resource provides all the detail—and code snippets—you’ll need to start manipulating and visualizing time-series data, and pays extra attention to sometimes tricky-to-handle datetime formats.
- How to create a full time-series workflow. Once you feel comfortable with the building blocks of time-series forecasting, you may want to implement them into an end-to-end project. Marco Peixeiro’s comprehensive guide will set you on the right path: it explains how to frame a time-series problem as a supervised learning problem, which then allows you to use any scikit-learn model.
- When deep learning meets time series. How can we apply the power of neural networks to forecasting tasks? Gabriele Orlandi explores the potential applications of cutting-edge models in the realm of time-series analysis, and shows how the latest research is shaking up classical approaches to prediction.
If history teaches us anything, it’s that TDS readers are always curious to learn more. We’re here to help, with a selection of must-reads on other topics:
- In our latest Author Spotlight, we were thrilled to chat with NLP expert Julia Turc about multimodal machine learning and how early-career practitioners should choose the right project (among other topics).
- Writer-developers, rejoice: our friends at Medium announced the arrival of code blocks with syntax highlighting!
- For their debut TDS post, Anna Arakelyan and coauthor Dmytro Karabash made a splash with a deep dive on a “rainbow method” for label encoding.
- If you wanted to learn more about the emerging field of TinyML but weren’t sure where to start, Rafael Tappe Maestro’s first TDS article (co-written with Nikolas Rieder) is an accessible and engaging introduction.
- Covering both the theory and practice of topic modeling, Lan Chu and Robert Jan Sokolewicz walk us through a full implementation of Latent Dirichlet Allocation (LDA).
- Switching between programming languages is rarely easy, which is why Madison Hunter’s tips on moving smoothly between Python and R are essential reading for data scientists.
Thank you, as always, for your support. If you’d like to make the biggest impact, consider becoming a Medium member.
Until the next Variable,
TDS Editors
