Case Studies

Discover some of our expertise through the case studies below

Showcasing our analytics and statistical solutions.

  • Home
  • Case Studies

Finance

Portfolio Rebalancing

Problem: The existing portfolio did not achieve optimal diversification. The portfolio comprised of indivisible assets which complicated the process of reallocating funds.

Solution: Developed a program designed to analyse and produce an optimised portfolio for each client. The program identifies which assets should be purchased or sold to achieve the desired level of diversification.

Tools used: Python, Excel

Automation

Problem: The portfolio, consisting of dozens of individual client portfolios, lacked a user-friendly interface for analysts to view the portfolio as a whole or to examine specific areas. Additionally, portfolio forecasting was nonexistent.

Solution: Developed a forecasting model which worked on any asset in this class, so as assets are purchased, they are automatically added to the portfolio and forecasted. All of this was fed into an interactive dashboard for analysts to use.

Tools used: Python, SQL, Power BI

e-Commerce

Online advertising optimisation

Problem: Return on advertising spend across all channels (e.g., Google, Facebook) was not properly tracked, leading to inefficient allocation of the advertising budget.

Solution: Collected and cleaned all relevant data, then analyzed key metrics such as CPC, CPA, and CTR to determine the optimal allocation of the advertising budget moving forward. Implemented ongoing data monitoring to allow for continuous adjustments and improvements to maximize performance.

Tools used: Python, Excel, Google Sheets, Power BI

Sales forecasting and inventory optimisation

Problem: Business struggled with striking a balance between ensuring there was enough stock to fulfil orders, whilst also maintaining a lean stock level for better cash flow.

Solution: Developed a demand forecasting model that accurately predicted future stock requirements. This allowed the business to place precise inventory orders, effectively preventing both under-ordering and over-ordering.

Tools used: Python, SQL, Excel, Streamlit

Sales matching algorithm

Problem: The business was unable to match recipients of marketing materials, such as catalogs, to actual product orders. As a result, the effectiveness of marketing spend remained unclear.

Solution: Developed a bespoke algorithm that leverages multiple variables to accurately match sales with catalog recipients. The algorithm achieved a 95% accuracy rate, providing valuable insights into the effectiveness of the marketing campaigns.

Tools used: Python, Streamlit

Research

Survey Analysis

Problem: A large national survey required detailed analysis across numerous subcategories.

Solution: Gained a thorough understanding of the research objectives and applied the appropriate statistical methods to accurately analyse the data. This ensured that any statistically significant relationships were identified.

Tools used: R, R-Studio, PPT

Gaming

Value betting website

Problem: New value betting website needed help with deciding which sports to focus on.

Solution: Creators of the site had over 1 million rows of betting and prediction data for over 30 sports. We had the task of finding which sports exhibited the least efficient betting markets, allowing the online software to be developed with these sports being targeted.

Tools used: Python, Excel

Strategy optimisation

Problem: Client had 3 years worth of betting history (500,000+ bets placed) using multiple strategies, but lacked a system for optimisation.

Solution: Backtested their current strategy, finding areas for improvement. Developed a staking strategy which was not currently being employed.

Tools used: Python, Excel

Custom tool development

Problem: Clients' daily betting routine was littered with manual processes which took time and allowed for the introduction of human error.

Solution: Gained a thorough technical understanding of their processes, while also grasping the underlying reasons behind them, not just the surface-level actions. Developed custom software which removed the chance of human error and saved up to 2 hours per day.

Tools used: Python, Streamlit

Education

Student Analytics

Problem: Education and early careers firm has 10k+ users, all of which need tailored recommendations.

Solution: Through the use of statistical analysis, we were able to categorise different students, whilst also assigning probabilities to different outcomes.

Tools used: Python, Excel

Other

Custom tool development - recommendation tool

Problem: When industries decline or significant job losses occur, there are excess workers within a region that need to find new employment opportunities.

Solution: Developed a tool using mathematical techniques (Euclidean distances) to identify the most similar jobs and industries based on required skill sets.

Tools used: R/R Studio, Python

  • Finance
  • E-Commerce
  • Research
  • Gaming
  • Education
  • Other

Finance

Portfolio Rebalancing

Problem: The existing portfolio did not achieve optimal diversification. The portfolio comprised of indivisible assets which complicated the process of reallocating funds.

Solution: Developed a program designed to analyse and produce an optimised portfolio for each client. The program identifies which assets should be purchased or sold to achieve the desired level of diversification.

Tools used: Python, Excel

Automation

Problem: The portfolio, consisting of dozens of individual client portfolios, lacked a user-friendly interface for analysts to view the portfolio as a whole or to examine specific areas. Additionally, portfolio forecasting was nonexistent.

Solution: Developed a forecasting model which worked on any asset in this class, so as assets are purchased, they are automatically added to the portfolio and forecasted. All of this was fed into an interactive dashboard for analysts to use.

Tools used: Python, SQL, Power BI

e-Commerce

Online advertising optimisation

Problem: Return on advertising spend across all channels (e.g., Google, Facebook) was not properly tracked, leading to inefficient allocation of the advertising budget.

Solution: Collected and cleaned all relevant data, then analyzed key metrics such as CPC, CPA, and CTR to determine the optimal allocation of the advertising budget moving forward. Implemented ongoing data monitoring to allow for continuous adjustments and improvements to maximize performance.

Tools used: Python, Excel, Google Sheets, Power BI

Sales forecasting and inventory optimisation

Problem: Business struggled with striking a balance between ensuring there was enough stock to fulfil orders, whilst also maintaining a lean stock level for better cash flow.

Solution: Developed a demand forecasting model that accurately predicted future stock requirements. This allowed the business to place precise inventory orders, effectively preventing both under-ordering and over-ordering.

Tools used: Python, SQL, Excel, Streamlit

Sales matching algorithm

Problem: The business was unable to match recipients of marketing materials, such as catalogs, to actual product orders. As a result, the effectiveness of marketing spend remained unclear.

Solution: Developed a bespoke algorithm that leverages multiple variables to accurately match sales with catalog recipients. The algorithm achieved a 95% accuracy rate, providing valuable insights into the effectiveness of the marketing campaigns.

Tools used: Python, Streamlit

Research

Survey Analysis

Problem: A large national survey required detailed analysis across numerous subcategories.

Solution: Gained a thorough understanding of the research objectives and applied the appropriate statistical methods to accurately analyse the data. This ensured that any statistically significant relationships were identified.

Tools used: R, R-Studio, PPT

Gaming

Value betting website

Problem: New value betting website needed help with deciding which sports to focus on.

Solution: Creators of the site had over 1 million rows of betting and prediction data for over 30 sports. We had the task of finding which sports exhibited the least efficient betting markets, allowing the online software to be developed with these sports being targeted.

Tools used: Python, Excel

Strategy optimisation

Problem: Client had 3 years worth of betting history (500,000+ bets placed) using multiple strategies, but lacked a system for optimisation.

Solution: Backtested their current strategy, finding areas for improvement. Developed a staking strategy which was not currently being employed.

Tools used: Python, Excel

Custom tool development

Problem: Clients' daily betting routine was littered with manual processes which took time and allowed for the introduction of human error.

Solution: Gained a thorough technical understanding of their processes, while also grasping the underlying reasons behind them, not just the surface-level actions. Developed custom software which removed the chance of human error and saved up to 2 hours per day.

Tools used: Python, Streamlit

Education

Student Analytics

Problem: Education and early careers firm has 10k+ users, all of which need tailored recommendations.

Solution: Through the use of statistical analysis, we were able to categorise different students, whilst also assigning probabilities to different outcomes.

Tools used: Python, Excel

Other

Custom tool development - recommendation tool

Problem: When industries decline or significant job losses occur, there are excess workers within a region that need to find new employment opportunities.

Solution: Developed a tool using mathematical techniques (Euclidean distances) to identify the most similar jobs and industries based on required skill sets.

Tools used: R/R Studio, Python

Unlock the power of data with RAMM Analytics, delivering customised solutions that drive results.

Linkedin

Other Pages

Copyright © 2024 – Rammanalytics. All rights reserved.

Scroll to Top