My Projects
Here are some of my projects i have been working on to enhance my skills in free time....
List of key projects which I have worked on, as a RTI Engineer in Thales:
- Entries Predictions - This project is about predictive modeling of entry data, which helps in forecasting the number of passengers entering the MTR station. We have tried to make predictions at different levels using different machine learning models. • Short Term that predict next 3 hours using past 2 years data to train LSTM. Mid Term that predict next 7 days using past 15 days entries data to train STLFArima. Overall Trend that predict next 3 hours using past 3 months entries, mid term prediction, edges data to train Keras NN model. • We also applied the same mid term and overall trend solution on waiting time data that help to predict waiting time of passengers at platform to catch the train at MTR stations.
- Exit Prediction Analysis - There are two algorithm created by Thales France & Canada to find out the destinations of passengers entering the station. Thales France’s algorithm support metro data and Thales Canada’s algorithm supports buses data. To perform the analysis, Biscay’s metro and bus trips data & GTFS timetable data were used and modified the above algorithm to examine their performance individually and later merged the results of both to see the impact. Also modified Thales France’s algorithm to adapt the bus data.
- Alarm Prediction – It is the FP-Growth model, which helps in detecting the avalanche of alarms occurring in the database and performs analysis on those alarm patterns. Also, applied association rules to predict alarms.
- Passenger Distribution - It is a Diffusion Metrics model, which helps in predicting the distribution of number of passengers in the MTR network. It predicts where and when the passenger entering the station will be.
- Anomalies Detection- This algorithm is designed to detect anomalies. If the number of passengers entering the station is more than the expected space capacity of the platform, it will raise an alert which will help the station operators to take necessary steps for crowd management.
- Naia Biscay Passenger Journey Analytics- Driven by the trend for improved passenger experience required by mass transit system operators, Thales has designed “Naia” as a Big data Analytics platform to help them to understand how passengers travel through their network in order to improve the quality of service offered.
- Passenger Flow Simulator Analysis- This is a simulator created by Thales France to visualize the flow of passengers in the MTR network. To perform the analysis, Hong Kong MTR trip data and timetable data of multiple lines were used to check the compatibility of the simulator to data from other countries.
- BEM Fare Allocation System (Bangkok) - It is Fare Allocation System for Bangkok Expressway & Metro company which maintain the blue line. This system used to manage ticketing data of BEM and automatic generation of financial reports.
- Naia Hong Kong Passenger Flow Analytics - It is generating 3 KPIs i.e Platform Crowding, Train Occupancy & Waiting Time and generating route followed by passenger in MTR network from origin to destination.
- BKKBLE - Blue Line Extension (Bangkok)
- BKKBLU- Blue Line Upgrade (Bangkok)
- SZL4- Shenzhen Line 4(Red Line)
- Delta.io vs Clickhouse Data Pipeline Comparison: Conducted a performance comparison of two different database technologies for extracting, transforming and loading large dataset.
- Auto-Tool to update IO Configurations in Excel: Developed an automated tool using python to modify macro-enabled excel files by comparing them with IO list configuration files. This tool streamlined the System Engineering process of updating Input/Output configuration for various equipment in MTR stations.
- NLPtoSQL - Build a model to convert Natural Language Queries to SQL queries using LLMs and Hugging Face.
- NLPtoDoc - Build a model to summarize given document based on question asked by user using LLMs and Hugging Face.
My Certifications
Here are some of the latest certifications i have been certified on.
My Skills
A RTI Engineer needs enormous range of skills and knowledge to improve the decision-making process.
Below are some skills that I have.
85
Python
80
Machine Learning
75
Deep Learning
90
Data Visualisation
85
SQL
75
Statistics
80
Data Structures and Algorithm
70
DevOps
90