Freelancers using Streamlit in DhakaFreelancers using Streamlit in Dhaka
ProjectsPeople
Istiak Ahmed Khan's avatar
Istiak Ahmed Khan
Dhaka, Bangladesh
Power BI Data Analyst + ML AI Automation Expert
5.0
Rating
102
Followers
Power BI Data Analyst + ML AI Automation Expert
Cover image for Education Program Analytics Dashboard –
Education Program Analytics Dashboard – Data Analytics Solution Handling large-scale program data without clear visibility can make impact measurement difficult. This dashboard is designed to give organizations a complete, real-time view of program performance, reach, and funding — all in one place. What You Get : A powerful, interactive dashboard that helps track beneficiaries, program outputs, regional performance, and donor contributions with clarity and precision. Key Capabilities: Monitor total beneficiaries and gender distribution Track program reach across provinces and sectors Analyze quarterly trends and growth patterns Evaluate donor funding allocation and impact Identify top-performing program categories and outputs Explore education level distribution and engagement For NGOs and large programs, data is critical for decision-making and reporting. This dashboard helps you measure impact, improve transparency, and optimize resource allocation — making your data meaningful and actionable. Perfect For : NGOs and non-profit organizations Government programs International development agencies Research and policy teams If you want a professional, insight-driven dashboard that clearly communicates impact and performance, I can create a customized solution tailored to your organization.
1
268
Cover image for 3D Molecular Intelligence: Accelerating Drug
3D Molecular Intelligence: Accelerating Drug Discovery through Predictive Analytics The Impact: This project addresses a critical bottleneck in pharmaceutical research: determining molecular solubility (logS). By replacing slow, expensive lab tests with high-precision machine learning, this system enables scientists to screen thousands of compounds in seconds, significantly reducing the cost and time required to bring new life-saving drugs to market. Drastic Cost Reduction: The predictive pipeline reduces early-stage experimental screening costs by 70–90%, allowing research teams to focus resources on the most promising drug candidates. High-Precision Forecasting: Utilizing a hybrid 3D feature engineering approach, the system achieves a remarkable 91.3% accuracy (R² score) in predicting solubility, providing a highly reliable alternative to physical testing. Accelerated R&D Cycles: By automating the identification of viable molecules, the tool dramatically shortens the "hit-to-lead" time in pharma and materials science, getting products to market faster. Empowering Researchers: I deployed a professional Streamlit dashboard featuring an interactive 3D molecular viewer. This allows non-technical chemists to visualize complex structures and make data-driven decisions without needing to write a single line of code
3
292
Cover image for Predictive Marketing Analytics: Optimizing Advertising
Predictive Marketing Analytics: Optimizing Advertising ROI 1. The Business Problem Companies often struggle to determine which marketing channels actually drive revenue. Without a data-driven approach, advertising budgets are often misallocated across platforms like TV, Radio, and Newspapers, leading to inefficient spending and missed sales targets. This project aimed to build a predictive model to quantify the relationship between multi-channel marketing spend and total sales. 2. Strategic Insights & Market Analysis Through a rigorous analysis of historical advertising data, I identified the specific drivers of revenue growth: Dominant Revenue Driver: TV advertising emerged as the most critical factor, showing a massive 0.9 correlation with sales. Efficiency Analysis: While Radio and Newspaper spending contributed to the marketing mix, their direct impact on sales was significantly lower (0.35 and 0.16 correlation, respectively), suggesting a need for budget reallocation. Predictive Power: My analysis revealed that 81.6% of the variance in sales can be explained by TV advertising spend alone, providing a highly reliable foundation for future budget forecasting. 3. Data-Driven Solution I developed a Linear Regression model to provide leadership with a mathematical framework for sales forecasting. Reliability: The model was validated using a 70/30 train-test split, ensuring it performs accurately on new, unseen market data. Accuracy: The system achieved a strong R-squared value of 0.79 on the test set, meaning it can accurately predict nearly 80% of sales fluctuations based on planned marketing spend. Error Management: I performed a detailed residual analysis to confirm that the model’s error terms were normally distributed, ensuring the reliability of the forecasted figures. 4. Business Impact Budget Optimization: Provided a clear mathematical equation (Sales=6.948+0.054×TV) that allows the marketing team to calculate the expected return on every dollar spent on TV advertising. Strategic Planning: Enabled the transition from "gut-feeling" marketing to precision budgeting, allowing the company to maximize ROI by prioritizing high-impact channels. Risk Mitigation: By identifying the variance that the model couldn't explain, I helped the business identify where external market factors might still influence sales, allowing for more conservative and realistic financial planning. Technical Stack Modeling: Simple Linear Regression, Statsmodels (OLS), Scikit-learn. Analytics: Python, Pandas, NumPy. Visualization: Seaborn, Matplotlib, 3D Scatter Plots
4
5
688
Cover image for Sentiment-Driven E-commerce Optimization: Amazon Review
Sentiment-Driven E-commerce Optimization: Amazon Review Analysis & Rating Prediction. Project Overview ✅ This project serves as a machine learning proof of concept designed to transform raw Amazon product reviews into actionable business insights. By automating the prediction of review ratings and analyzing customer sentiment, the system enables brands to optimize product listings, proactively address customer pain points, and drive higher conversion rates. Process ✅ I developed an end-to-end pipeline covering data acquisition, complex text processing, and model deployment. Automated Data Scraping, Integrated the Apify API to extract real-time customer feedback directly from Amazon product URLs. I configured the scraper to handle up to 100 reviews per run, capturing critical metadata including rating scores, review descriptions, and verified purchase status. Data Refinement & Feature Engineering: Cleaned a dataset of approximately 1,944 reviews by removing noise (punctuation/symbols) and stop words using NLTK. I implemented TF-IDF Vectorization to convert text into numerical features and applied SMOTE (Synthetic Minority Over-sampling Technique) to address class imbalance, ensuring the model could accurately predict rare negative reviews. Model Benchmarking ✅ Developed and compared three distinct architectures to identify the most robust predictor: Naive Bayes: High-speed probabilistic classification. Support Vector Classifier (SVC): Optimized for high-dimensional text data. Neural Network (MLPClassifier): To capture complex semantic patterns. Web App Deployment: Built a dedicated Streamlit dashboard that allows non-technical stakeholders to input raw review text and receive instant rating predictions with confidence scores. Technical Stack✅ Languages & Tools: Python, Apify Client. ML & NLP Libraries: Scikit-learn (SVC, Naive Bayes, MLP), NLTK (Tokenization, Stopwords), Imbalanced-learn (SMOTE). Deployment: Streamlit, Joblib (Model Serialization). Visualization: Plotly, WordCloud, Matplotlib. Key Results ✅ Achieved a peak accuracy of 95.27% using the Neural Network model, with the SVC model following closely at 94.46%. Developed sentiment-based feedback loops within the app: high ratings (4-5 stars) trigger positive marketing recommendations, while low ratings (1-2 stars) alert teams to address product issues like battery life or build quality. Enabled real-time competitive analysis by providing a user-friendly interface for cross-functional marketing and product development teams to audit customer sentiment at scale.
2
5
726
Streamlit(5)
Shanzia Shabnom Mithun's avatar
Shanzia Shabnom Mithun
Dhaka, Bangladesh
I build AI systems and document them clearly
New to Contra
I build AI systems and document them clearly
Cover image for API-Triggered AI Dispatch Automation Pipeline Built
API-Triggered AI Dispatch Automation Pipeline Built a live automated dispatch system that receives incoming service requests, analyzes them with an LLM, and routes field technicians intelligently, without human triage. The pipeline uses FastAPI as the backend engine, OpenAI for decision logic, and Streamlit for real-time monitoring. Designed and deployed solo from architecture through production. The system is live and publicly accessible.
0
46
Cover image for Automated Bilingual Document Intelligence System
Automated Bilingual Document Intelligence System (RAG + LLM) Built a retrieval-augmented generation system that allows users to query a specialized document corpus in both English and Bengali. The pipeline ingests source documents, indexes them with vector embeddings, and returns accurate, context-aware answers via an LLM. Designed for real-world use in water, sanitation, and hygiene (WASH) program contexts. Deployed live on Streamlit Cloud.
0
29
Cover image for Automated Medical Image Analysis Pipeline
Automated Medical Image Analysis Pipeline with Domain Generalization Developed a deep learning pipeline that classifies retinal fundus images across multiple disease categories, trained to generalize across unseen datasets using leave-one-domain-out evaluation. Built with PyTorch and deployed as an interactive Streamlit application. Part of active ML research exploring cross-dataset robustness for medical imaging. Combines research-grade methodology with production deployment.
0
38
Cover image for Automated Semantic Search and Matching
Automated Semantic Search and Matching Engine (FAISS + SBERT) Built a semantic product recommendation pipeline using FAISS for vector indexing and Sentence-BERT for generating dense embeddings. The system matches user queries to relevant products based on meaning not just keyword overlap, enabling highly relevant recommendations at scale. Fully deployed and live, demonstrating real-world vector database architecture on a production workload.
0
40
Streamlit(1)