Welcome to my Portfolio

Topic: Analysis of Global Video Games Market to decide business strategy for a game retail business

Insights derived: Video games releases and sales have seen a decline since the 2010s, which is an opposite trend from mobile games.

Also, North America is the biggest market and action is the most popular game genre.

Programs used: Microsoft Excel with Analysis Toolpack, Pivot Charts etc.

Topic: Analysis of regional expansion of a clothing retail business in post-COVID period with final recommendations

Insights derived: Using a weighted evaluation scorecard across COVID-19 and business metrics, Singapore is the best regional country to expand to in a post-Covid world, followed by Brunei then Thailand.

Programs used: MS SQL (database and ETL process), Power Query (misc. cleaning), Microsoft Excel with Analysis Toolpack, Pivot Charts

Topic: Analysis of Olist's Brazilian e-commerce operational data for sales trend, geographical spread etc. with suggestions

Insights derived and recommendations:
Sales has been below trendline from May 2018 to August 2018.
Recommended to investigate, especially the consistently moderate volume of 1 star reviews, since this signals operational flaws

Trends are developing towards card payments and health and beauty product categories.
Recommended to work with banks to promote card usage, and market Olist as an attractive platform for health and beauty products

Programs used: Power Query, Power BI

Please refer to the original jupyter notebook code at: Github / Mirror

Topic: Supervised machine learning of Diabetes symptoms' relationships to Diabetes risk, leading to feature importance and predicative machine learning from custom inputs of symptoms

Insights derived and recommendations:
It seems that symptoms of excess drinking and excess urination are prominent in patients with diabetes. A combination of other factors can lead to more accurate prediction of diabetes, which the prediction model can achieve.

Recommended to test diabetes risk level with the prediction model, and to see a local doctor if one is determined to be at high risk of contracting diabetes

Programs used: Microsoft Excel, Jupyter Notebook (Python and Machine Learning)