Python is the most dynamic instrument for data scientists and its results are extremely easy to integrate into business processes. Unlike most other analytical and statistical software, Python is free of charge and it is used 4-5 times more often in data analysis.
Please note that the workshop is going to be very useful even for those who have never used Python before since all steps will be shown at the projector and you will receive all the materials.
At this workshop you will be able to learn about:
–Anaconda and modules of Python; how to upload users’ transactions and aggregate the data
–Behavior of anomalous users, how to check for outliers with the help of multivariate normality test
–Data analysis with various tests usage
–Clustering algorithms (k-means, dbscan), dimensionality reduction (t-sne, PCA) and other methods of unsupervised learning
–Interpretation of results and visualization
And you will get lots of useful tips from the speaker ; )
Tentative schedule:
11:00 – 12:30 Intro of the speaker and the event agenda. Homework discussion
12:30 – 13: 00 Download transactions by users. Aggregate them
to the user level in Python, visualize the results
13:00 – 13: 30 Lunch break
13:30 – 14: 30 Finding anomalous transactions and anomalous users
14:30 – 15:30 User segmentation using K-means. Determining the optimal number of clusters. Interpretation of results. Results visualization.
15:30 – 16:00 Q&A
Our speaker is Kyryl Iurchenko, Data Science Architect at GlobalLogic, Financial Risk Manager (Certified by the Global Association of Risk Professionals), Master’s in Econometric Analysis and Machine Learning from Kyiv National Taras Shevchenko University.
Requirements:
– Necessary: laptop with installed Anaconda (Links and manual will be sent)
– Optionally (for those who are beginners in Python): make a small task in Python before the event (will be sent to registered guests later on)
Language: Ukrainian
Place: Kyiv School of Economics, Dmytrivs`ka st., 92-94