Omnis tempus datum

Every moment is a data point

Blog documenting the tinkerings and helpful tips I have accumulated in the area of data science and machine learning, with some flavor of chemical science sprinkled sporadically.

The posts are meant to be a reference for my future self or whoever wishes to start in the data science and ML space.

Posts

• Anamoly detection using autoencoders

  Vanilla autoencoder to learn a distributed representation of MNIST data, generate new instances, and estimate anomalies

  Mar 18, 2021

• Implement neural network from scratch for binary classification

  Build a single-layer neural network model for binary data classification

  Feb 5, 2021

• t-SNE and UMAP - Effect of initialization on the dimensionality reduction

  Feb 3, 2021

• Bayesian optimisation implementation

  Walkthrough on implementing a modular bayesian optimization routine

  Jan 27, 2021

• Estimating prediction confidence through dropout

  Jan 11, 2021

• Solving the birthday problem using simple counting

  Counting trials instead of using a PnC formula

  Jan 10, 2021

• Central limit theorem

  Visual proof of the central limit theorem

  Dec 21, 2020

• Relational analysis of spices used in Indian cuisine

  Analysis to understand which spices are frequently used in the Indian cuisine and explore the connection between them

  Dec 9, 2020

• Network analysis hands-on

  Nov 8, 2020

• Transfer learning walkthrough using Pytorch

  Implement transfer learning using pre-trained CNN

  Oct 19, 2020

• Get SMILES from PubChem using DASK

  Use parallelized DASK methods to find SMILE and InChI keys for molecules

  Sep 18, 2020

• Analyze Bollywood movie ratings (1950-2020)

  Analyze IMDb Bollywood movie data to find average ratings and number of bollywood movie releases over time.

  Aug 23, 2020

• Classification-based algorithms walkthrough

  Look at different types of classification types and metric implemented in scikit-learn by building a MNIST digit classifier

  Aug 12, 2020

• S&P 500 analysis using beautifulsoup and pandas

  Notebook that uses `BeautifulSoup` and `yfinance` api to accquire a list of S&P 500 companies and plot their market capital and year-to-date return

  Aug 1, 2020

• Plotting surface in matplotlib

  Simple notebook looking at meshes in matplotlib

  Jul 18, 2020