• Employee performance trend analysis in MySQL:
  Purpose: This project focused on identifying underperforming field force employees by analyzing their sales efficiency using Gross Profit (GP) leverage. GP leverage is defined as the ratio of gross profit generated to promotional expenses incurred by individual sales personnel. The objective was not only to detect low performers but also to diagnose underlying performance bottlenecks through targeted interviews.
  Data & Methodology: 

• The analysis was conducted using a panel dataset titled “consolidated_gp”, which contained monthly data (July–December 2022) for each field force employee, including unique ID, region, sales, gross profit (GP), promotional expenses, and calculated GP leverage.

• Using MySQL, the dataset was explored, cleaned, and structured. Date formats were standardized using built-in functions, and SQL queries were employed to filter and extract relevant subsets of data.

• To identify underperformers, field force employees were ranked based on GP leverage for December 2022. The bottom 300 employees with the lowest GP leverage were selected as the primary focus group. The maximum GP leverage within this group was used as a benchmark threshold for further filtering.

• A longitudinal analysis was then conducted on these 300 employees across the six-month period (July–December 2022). Quarterly segmentation was applied to compute the average GP leverage for Q3 (July–September) and Q4 (October–December) 2022 for each individual.

• To enhance interpretability, a performance classification framework was developed and implemented directly in MySQL using conditional logic. Employees were categorized as follows based on their average GP leverage: Poor: 0–1; Needs Improvement: 1–3; Meets Requirement: 3–4; Exceeds Requirement: 4–6; Excellent: ≥6

• This classification enabled comparative analysis of performance trends between Q3 and Q4. Particular attention was given to employees whose performance declined significantly (e.g., from “Meets Requirement” in Q3 to “Poor” in Q4), indicating potential operational or capability gaps.

• Accomplishment:

• The analysis revealed a subset of employees with noticeable performance deterioration over time. These individuals were shortlisted for further investigation through structured phone interviews.

• The qualitative insights gathered from these interviews highlighted several root causes of underperformance, including:

• Inadequate product knowledge and training
   

• Weak engagement and communication with target customers
   

• Misalignment between assigned sales targets and market conditions

• These findings were consolidated and presented to management, along with actionable recommendations. The project contributed to informed decision-making by enabling targeted interventions such as training programs, performance recalibration, and improved field force management strategies.

• 1+ other use case solutions in MySQL

• A Machine Learning Project on Classifying Happy or Unhappy People

• Purpose: To find the factors that associate with happiness.

• Workings: Removing unnecessary text from all 31 columns, encoding categorical variables using one-hot, dummy, and ordinal encoding, selecting features using the Boruta feature selection method, selecting training data, building models, creating classification reports and confusion matrices, plotting ROC curves, selecting the best model based on F1 score and AUC.

• Accomplishment: Successfully achieved 93% accuracy using the Gradient Boosting model in identifying the factors associated with happiness.

• 2+ other machine learning projects, 1+ other data wrangling project in Python

• Tableau Dashboard-1
  Purpose: Creating a dashboard in Tableau for visualizing pharmaceutical industry information as required by the management.
  Workings: Collecting data, cleaning data, creating table for Tableau connection, analysing the data as per requirements, and showing the analysis by using bar chart, bubble chart, and pie chart. 
  Accomplishment: Assisting the management in taking quick decisions by looking at the dashboard.

• 1 other Tableau dashboard 9+ other Tableau charts 

• Forecasting of Pharmaceutical Sales
  Purpose: To forecast the sales of Pharmaceutical Company "X" for the next 24 months by analyzing seasonality using R programming.
  Workings: Gathering 111 months of sales data, converting it into a time series, plotting ACF and PACF, using ggsubseries plot to identify seasonality, performing Box Cox transformation and taking the first difference, selecting the order of ARIMA, fitting ARIMA and ETS models, performing classical and STL decomposition, fitting ARIMA and STL models on decomposed data, selecting the best model, and forecasting sales for the next two years. 

• Accomplishment: Successfully analyzed the seasonality of Pharmaceutical sales in Bangladesh and made accurate sales forecasts for the next 24 months.

• 1+ other time series forecasting of S&P 500 and 3 valuation projects on Square Pharma, Grameenphone and Heidelberg Cement. 

• Power BI Dashboard-1:

      Purpose: Creating a dashboard in Power BI for visualizing business information        as required by the management.

      Workings: Collecting data, cleaning data, creating table for PBI connection,            analysing the data as per requirements and showing the info by using stacked        bar chart, donut chart, line chart, funnel chart, combo chart, and waterfall                chart.

      Accomplishment: Assisting the management in taking quick decisions by                  looking at the dashboard.

• 5+ other Power BI dashboards.

• Analytical Insights on Brands Performance:
  Purpose: Finding out the root causes of BDT 45 million less gross profit generation by top ten brands.
  Workings: Analyzing & visualizing the gross profit trend in tableau, conducting interviews, rating the causes based on impacts, and showing the insights on a spider chart.
  Accomplishment: Execution of action plans by the management to regain lost profits.

• 9+ other Business Analysis reports undertaken for management's requirement.

• Principal Component Analysis:
  Purpose: Achieving the optimum number of principal components from the iris data. This project was undertaken as course assignment.
  Workings: Finding variance co-variance matrix, finding eigen values & eigen vectors, evaluating principal components, and plotting the principal components and their associated variance explanation in scatter plot.
  Accomplishment: Learning of finding the optimum number of principal components and plotting the PCAs in scatter plot. 

• 2+ other multivariate projects regarding regression analysis and econometric analysis. 

• Statistical Inference Analysis on Medical Condition:
  Purpose: Hypothesis testing of mean level of cholesterol and creating 95% confidence interval. This project was undertaken as course assignment.
  Workings: Data manipulation, missing data finding, setting target and independent variables, plotting the correlations using seaborn module, conducting t test, and creating 95% confidence interval.
  Accomplishment: Learning of conducting statistical inference analysis by using python. 

• 1+ other hypothesis testing projects. 

• Detecting Outliers by using Box Plot in SPSS
  Purpose: Creating a tutorial on creating box-and-whisker plot in SPSS to detect outliers in a dataset.
  Workings: Uploading a real-world dataset into SSPSS, creating Box Plot is SPSS from graphboard template chooser, interpreting the result, and recording the tutorial in own voice.
  Accomplishment: Learning of detecting outliers form real world dataset.

• Cleaning Dirty Data in Excel
  Purpose: Cleaning dirty data downloaded from company software and turning it to a cleaned dataset. This project was undertaken as management requirements.
  Workings: Copying & pasting raw data sheet to another sheet, using logical statement to fill data in a single column, value pasting database in another sheet to clear formulas, designing table with the database, cleaning blanks and residuals, using logical statement to categorize customer type, using double v-lookup technique to group overdue ages, using conditional formatting to highlight long aged groups.
  Accomplishment: Saved 2 days of man-day of my department by using advanced logics. 

• 2 other Excel dashboards, and 5+ other excel data cleaning and analysis projects that consists advanced pivot table, graphs and charts, descriptive statistics, and formulas.