Build Models to predict sales, usedARIMA,LSTMandProphet,RandomForest, andGCPAutoml. build several models for different clients data is different and pick the best model working for their set of data to predict sales of their businesses.

Perform all different methods pre processing for different data so the accuracy of models will improve

Created Apis usingflaskfor the accessibility of models.

Analysis of all the Emails sent to customers and their actions. find entity in emails withSpacy NERand recommend words to improve subject lines by finding the closest word usingcosine similarity.

Created ahypothesismodel based on the business plan of the company to test out that if that current approach will be profitable or not for the organization.

NLP Based chatbot for the website using Universal encoder and cosine similarity Collected customers feedback data, pre-processed it, Then trained the model for intent Classi cation and applied parameter tuning to improve the accuracy.

For accessibility of data in understandable format created an automation tool which fetch the data from our sql server and update the google doc sheet.SQL queries are set based on required output and python script automate the whole task

Created a Machine learning tool which take the csv data and perform all the cleaning and pre processing steps with the selection of functions only. In advancement of the tool it can perform cleanup operation for the NLP tasks and also handle the big data , used Pyspark to work with the big data.

Generate graphs to get insights of data. send column data in format based on users requirement and on frontend it is visualize by using angulars high chars