Nirav Prajapati

interestingly, around 40% of the effort could be saved by implementing smart validation rules at data entry stage. Preventing incorrect or inconsistent data from entering the system in the first place can drastically reduce the time spent on cleaning messy datasets later. A small investiment in prevention can lead to huge gains in productivity and data quality.

10

2 Comments

Importance of Dashboard as a tool for Data Analyst. Dashboardis a useful component in a data analyst’s toolkit, and an essential skill in a data analyst’s skillset. The term ‘dashboard’ comes from the automotive industry where car designers have put the most important gauges and other display information, such as engine oil temperature, current speed, current RPM, amount of fuel left, and so on, in a handy graphical display that is easy for the driver to view and understand. Originally these displays were analog, but many are now digital and use varying forms of visualization, including digital meters and mini graphs.  As a data analyst, you can take that same idea and apply it to a dashboard in a data analysis application; It is highly important to put key business information in one place, in the form of graphical displays, to make it easier for the viewer to understand them. Dashboards can take this a step further by also allowing the user to interact with the dashboard and modify exactly what information they see by using tools supplied on the dashboard. Users of a dashboard therefore not only get a consolidated visualization of their business data and key performance indicators (or KPIs), but also get a controllable self-service business intelligence (BI) interface through the use of filters, which enable them to control precisely what information they see. Benefitsof Dashboard Includes the following It offers insights into your key data. It can alert you to patterns and trends in your data. It provides an interactive experience for the user; allowing them to filter what data they see. It also provides a centralized and consolidated view of business data. A dashboard can be a very useful tool in areas of the business such as fiscal forecasting and reporting, project management, executive reporting, human resources, customer service, helpdesk issue tracking, healthcare monitoring, call center analytics, social media marketing, and many more. For a budding data analyst, dashboards can be a vital skill to add their arsenal, as the majority of employers see the dashboarding skill as a ‘must have’ rather than a ‘nice to have’. 

1

🎬 Power BI Cinema Cluster Dashboard | Weekly Performance Comparison Proud to share this Power BI dashboard built for analyzing cinema-wise performance over two weeks (Wk1 vs Wk2) in XYZ City, focused on KPIs like: 📊 Key Performance Metrics Tracked: 🎟 Admits & Admits Share – Tracking audience footfalls and market share 🎬 Shows Count – Monitoring programming frequency 💰 ATP (Average Ticket Price) – Pricing insights by cinema 💸 GBOC Share – Gross Box Office Contribution by cinema 🏆 Identified leading cinema across weeks based on admits & revenue contribution 🔍 Insights Derived: ✅ AA Cinema maintained market leadership with consistent share in admits & shows ✅ DD Cinemas improved in Wk2 with a rise in share and strong presence in specific locations 📉 CC Cinema underperformed, indicating scope for content strategy, timing, or programming optimization 🎥 Movie-level breakdown helped identify high-ATP movies and underperformers. 📌 Visuals include stacked bar charts for easier comparison of ATP, GBOC, Admits & Shows by cinema over the two weeks. 💡 Tools Used: Power BI, advanced visuals, KPI cards, slicers, DAX, and performance benchmarking logic. This dashboard offers data-driven direction for programming, pricing, and operational strategy in cinema clusters. #PowerBI #CinemaAnalytics #BoxOfficeInsights #DataVisualization #DataAnalytics #FilmBusiness #PowerBIProjects #BusinessIntelligence #MediaAnalytics #MoviePerformance #DAX

3

FMCG Festive Data: A first-party reservoir ready to burst? Analysts estimate FMCG brands are sitting on a ₹500 crore “goldmine” if they treat first-party data not just as an activation pipe for ads, but as a true intelligence engine. Read the full story by Shantanu David. Limk in comment. Shradha Agarwal, Grapes Worldwide | Varun Mohan, MiQ | Saloni Jain, Plus91Labs | Swagatika Das, Nat Habit | Anil K Pandit, Publicis Media #e4m #FMCG #FirstPartyData #DataIntelligence #FestiveMarketing #ConsumerInsights

9

1 Comment

Analytics doesn’t create value but decisions do. 6+ months into a Business Analytics role in FMCG, after my MBA from IIM Kashipur, I’ve realised this early on: Good analytics rarely shouts. It quietly helps leaders take better decisions. What the role has taught me so far: • Data is often incomplete but business still moves • The hardest part isn’t building reports, it’s aligning stakeholders • One clear insight for sales / supply / ops beats 10 detailed slides • Understanding the business rhythm matters as much as technical depth What my MBA & IIM experience truly helped with: • Thinking in trade-offs, not perfect answers • Staying calm in ambiguity • Asking why this matters before how to present it • Keeping business impact at the centre of analysis Still learning. Still refining my judgement. But growing more comfortable with decision-led, practical analytics every day. In FMCG especially, the best insights often come from conversations, not dashboards. #BusinessAnalytics #FMCG #DecisionMaking #Leadership #IIMKashipur #LearningOnTheJob

94

📊🚀 Blinkit-Style Power BI Interview Questions Every Analyst Should Practice It usually starts with a dashboard. Then comes the real test 👀 Why this metric? How would you improve performance? What story does the data actually tell? This Blinkit Power BI Interview Q&A PDF covers questions that interviewers use to judge how you think, not just what you know 👇 🔹 Real-world business scenarios & case-style questions 🔹 KPI selection & metric justification logic 📈 🔹 Dashboard design & performance optimization 🔹 Handling messy data & edge cases 🔹 Translating insights into business decisions 💡 Perfect for candidates preparing for Data Analyst / BI / Power BI roles who want to move from tool usage to analytical thinking. 📌 Shared for learning & awareness. PDF is not original content. 📲 Join my WhatsApp channel for daily SQL & analytics interview prep : https://lnkd.in/gbsnzzKb 🔁 Follow Ajay Yadav for more Power BI, SQL & Data Analytics interview prep #powerbi #dataanalytics #interviewquestions #blinkit #businessintelligence #datascience

90

5 Comments

🎥 Project Showcase: Diabetes Nutrition & Glycemic Control Dashboard in Power BI 🩺📊 This dashboard transforms nutrition & glycemic data into actionable insights: ✅ Daily calories & nutrient balance ✅ Glycemic index trends across meals ✅ Group-wise analysis (Diabetic vs Non-Diabetic, Active vs Not Active) ✅ Veg vs Non-Veg diet comparisons 💡 Helps patients, doctors, and researchers make smarter decisions for better health outcomes. #PowerBI #HealthcareAnalytics #DiabetesCare #DataVisualization #Tech #Diabetes #Death #Recovery #BI #Diet #Research Microsoft Power Platform

3

Bronze stores everything. Silver decides what's worth keeping. Silver layer is where I: ● Remove true duplicates ● Handle nulls intentionally ● Standardize formats ● Apply business rules ● Validate data quality But NOT where I: ● Aggregate ● Join across domains ● Create business metrics Silver = clean, validated, single-source-of-truth Think of it like a kitchen: ● Bronze = raw ingredients from the store (still in packaging) ● Silver = washed, chopped, prepped (ready to cook) ● Gold = finished dishes (ready to serve) The task: Clean the orders table. Potential Problems: ● orders with null customer IDs ● Dates stored as strings ● Status values with inconsistent casing Let's fix them. Recap from last week below:

8

Stop guessing and start scaling your FMCG distribution. 📈 A modern Distributor Management System (DMS) provides the real-time visibility needed to lower DSO, optimize inventory, and improve distributor cash flow. Eliminate the blind spots in your supply chain and drive data-backed growth. 🚀 Read more: https://lnkd.in/gK2BBB9a #FMCG #SupplyChain #DMS #Automation #Distribution

9

To make individual predictions from XGBoost Regressor more interpretable, I integrated SHAP values and visualized them using the Waterfall Plot. In a recent project, I built an XGBoost regression model to forecast product-level sales for SuperKart stores, aimed at helping optimize inventory and regional planning. While the model performed well in capturing complex store–product interactions, the next challenge was explaining why it made a particular prediction for a given input. Predictions from complex ML models like XGBoost often deliver high accuracy but are rarely straightforward to interpret. Understanding why a model made a certain prediction is critical, especially when deploying models in business settings where decisions must be explainable. To make individual predictions more interpretable, I integrated SHAP (𝗦𝗛𝗮𝗽𝗹𝗲𝘆 𝗔𝗱𝗱𝗶𝘁𝗶𝘃𝗲 𝗲𝘅𝗣𝗹𝗮𝗻𝗮𝘁𝗶𝗼𝗻𝘀) and visualized them using the 𝗪𝗮𝘁𝗲𝗿𝗳𝗮𝗹𝗹 𝗣𝗹𝗼𝘁. SHAP decomposes a model’s prediction into additive feature contributions: 𝗣𝗿𝗲𝗱𝗶𝗰𝘁𝗶𝗼𝗻 = 𝗕𝗮𝘀𝗲 𝗩𝗮𝗹𝘂𝗲 + Σ(𝗳𝗲𝗮𝘁𝘂𝗿𝗲 𝗲𝗳𝗳𝗲𝗰𝘁𝘀) The base value represents the model’s average prediction, and each SHAP value quantifies how much a feature pushes the output above or below that baseline. The Waterfall Plot visualizes this process clearly: * The base value serves as the starting point. * Red bars indicate features pushing the prediction higher. * Blue bars indicate features pushing it lower.  The bar length represents the magnitude of each feature’s effect, showing how the model moved from the average to the final predicted value. However, 𝗜𝘁’𝘀 𝗶𝗺𝗽𝗼𝗿𝘁𝗮𝗻𝘁 𝘁𝗼 𝘀𝗲𝘁 𝗰𝗹𝗲𝗮𝗿 𝗲𝘅𝗽𝗲𝗰𝘁𝗮𝘁𝗶𝗼𝗻𝘀: SHAP values explain how input features influenced a specific prediction based on the model’s learned relationships they do not establish causality or guarantee business impact. Incorporating such explainability bridges the gap between performance and trust—making complex models transparent, auditable, and ready for real-world decision-making. P.S Will share a detailed overview of the project soon! ♻️ If you find this helpful - do like & repost. I'm Varun Sagar Theegala - follow along as I share my daily learnings, reflections, and experiences from my 3 years+ journey in analytics, data science, and AI. Tagging individuals who help many learners with Data Science/ML upskilling & preparation. Hope this adds value to your community! - Venkata Naga Sai Kumar Bysani Tajamul Khan Andres Vourakis Timur Bikmukhametov, PhD Karun Thankachan Nirmal Budhathoki #DataScience #MachineLearning #ModelTraining #SHAPValues #ModelExplainability

30

6 Comments

Southwest Bengaluru spends most on FMCG per kg !! I just got this report " FMCG PULSE June 2025" from Ramakrishnan of Kantar. Here are my learnings: 1. In the first quarter of 2025, in home consumption grew at 3.5 %, the lowest since 2022. This 3.5 % is split as 4.4% in Urban India and 2.7 % in Rural India. 2. The 22 listed FMCG companies grew urban volume only at 2.1 %. Rest of brands grew at 3.8 % while unbranded is at 8.4 %. Clear downgrading happening. 3. In rural , listed brands are growing at 5.1%, the only bright spot for listed FMCG. 4.The price paid per kilogram for FMCG has gone up by Rs 8 in the last one year, Q O Q, despite downgrading. 5.Washing liquids is the fastest growing FMCG category at 2.7 X growth since 2023. After this, batters and ready to cook foods are next. 6. Ready to eat is losing relevance. consumers want speed but willl not give up freshness 7. Cornflakes has seen losses over the last two years and struggling. The cornflakes leadership over many years and poor innovation are reasons for this. 8. A consumer has 156 shopping trips per annum , which means the consumer is buying every 56 hours. 9. Consumers made 44 out of home trips annually. They are buying 4 additional packs bringing total packs bought to 176. 10. OOH FMCG has done poorly , a 4 pc growth. 11. Some brands have done well in some geographies. Sting did well in the top 6 metros, Milk Bikis single handedly contributed to the growth of biscuits in the top 6 metros, Sprite grew in large and small metro towns. 12. Kantar has taken the 9 big cities and broken them to 157 micro clusters. The first time at least I know this was done was when Vikesh Chandna did this at Nokia in 2010. 13. In Seven clusters, consumer more than 200 kgs of FMCG, six are in Delhi, one in Hyderabad. 14. Average FMCG spend is Rs 19,000 per annum, West Delhi (Tilak Nagar, Janakpuri,Sagarpur and Vikaspuri) is Rs 33,000 15. At Rs 227 per kg spent on FMCG, Southwest Bengaluru comprising RR Nagar, Mysore Road and Kengeri is the most premium. 16. Bengaluru spends Rs 211 per kg of FMCG, which is far higher than the national average of Rs 195. Only one cluster in Bengaluru is below Rs 200 per kg bought. Bengaluru spends on ready to cook and ready to eat, talcum powder and deodorants. They don't spend much on hair oils. Good report Enjoy shiv www.shivshivakumar.com Kantar K Ramakrishnan Weekly learnings 29/25

310

27 Comments

🚀 Exciting news for all data enthusiasts! 📊 Embracing the power of data science tools not only amplifies your analysis capabilities but can also foster a culture of innovation within your team. Rather than viewing these tools as mere instruments, consider them as catalysts for uncovering hidden patterns and generating valuable insights. 💡 Here's the twist – while we often focus on the technical functionalities of these tools, let's not overlook their potential to spark creativity. By encouraging a playful and exploratory approach to utilizing data tools, we pave the way for groundbreaking discoveries and unconventional solutions. 🔍 Challenge yourself to think beyond conventional applications and leverage data science tools as a canvas for experimentation. Who knows, your next data visualization or predictive model could be the spark that ignites a new era of innovation in your organization! #DataScience #Innovation #Creativity #DataTools #Analytics #TechForward

1

𝐎𝐧𝐞 𝐨𝐯𝐞𝐫. 𝐒𝐢𝐱 𝐛𝐚𝐥𝐥𝐬. 𝐒𝐢𝐱 𝐬𝐢𝐱𝐞𝐬. Yuvraj Singh didn’t just hit boundaries that night, he changed the fundamentals of T20 cricket. From cautious play to fearless aggression. From survival to domination. 𝐓𝐡𝐚𝐭 𝐦𝐢𝐧𝐝𝐬𝐞𝐭 𝐬𝐡𝐢𝐟𝐭 𝐬𝐡𝐚𝐩𝐞𝐝 𝐭𝐡𝐞 𝐰𝐚𝐲 𝐈𝐧𝐝𝐢𝐚 𝐩𝐥𝐚𝐲𝐬 𝐓20𝐬 𝐭𝐨𝐝𝐚𝐲, 𝐛𝐨𝐥𝐝, 𝐝𝐞𝐜𝐢𝐬𝐢𝐯𝐞, 𝐚𝐧𝐝 𝐛𝐮𝐢𝐥𝐭 𝐟𝐨𝐫 𝐢𝐦𝐩𝐚𝐜𝐭. Business leaders face a similar inflection point. For years, we chased growth. Then reality hit: 𝐍𝐨𝐭 𝐚𝐥𝐥 𝐠𝐫𝐨𝐰𝐭𝐡 𝐢𝐬 𝐠𝐨𝐨𝐝 𝐠𝐫𝐨𝐰𝐭𝐡. That’s where two conversations collide: RGM (Revenue Growth Management) — 𝐇𝐨𝐰 𝐝𝐨 𝐰𝐞 𝐠𝐫𝐨𝐰? PGM (Profitable Growth Management) — 𝐖𝐚𝐬 𝐭𝐡𝐚𝐭 𝐠𝐫𝐨𝐰𝐭𝐡 𝐰𝐨𝐫𝐭𝐡 𝐢𝐭? Over the next few weeks, I’ll share 5 focused insights on connecting these worlds, turning speed into sustainability and ambition into value. Because the future belongs to leaders who ask both: “𝐇𝐨𝐰 𝐦𝐮𝐜𝐡 𝐝𝐢𝐝 𝐰𝐞 𝐠𝐫𝐨𝐰?” 𝐚𝐧𝐝 “𝐖𝐚𝐬 𝐢𝐭 𝐰𝐨𝐫𝐭𝐡 𝐢𝐭?” Stay tuned for: 1. Growth ≠ Value 2. Pricing for Margin 3. GTM vs Discipline 4. When RGM must graduate into PGM 5. The Integrated Playbook Your turn: In your organisation today, which conversation dominates more, 𝐫𝐞𝐯𝐞𝐧𝐮𝐞 𝐨𝐫 𝐩𝐫𝐨𝐟𝐢𝐭? #RGMRedefined #ProfitableGrowth #StructureVsStorm #RGM #GrowthStrategy #BusinessLeadership #RevenueManagement #IndiaBusiness #ConsumerEconomy

184

6 Comments

Data Governance is Not Optional.   Business mein hum aksar data issues sunte hain — duplicate entries, missing fields, inconsistent formats. Abhi recent news mein bhi aise hi issues voter lists ke around saamne aaye.   No matter which side of the debate you stand on, this brings forward a crucial lesson: ➡️ Data governance is not optional. Quality checks are the backbone of trust.   Some patterns highlighted are duplicate records, invalid addresses, bulk entries, are not very different from what we see in enterprise systems when data ingestion, validation, and monitoring are not robust.   🔑 Key takeaways for us : ✅Validation at Source: Garbage in = garbage out. ✅Remove duplicates: Keep only one clean record. ✅Be transparent: People trust data when it is checked and reviewed. ✅Decide who owns the data, who checks it, and who corrects it.   Ultimately, data is not just numbers on a sheet; it is trust.   Whether in a company data/dashboard or a voter list, people expect reliability.

11

🎉 From the desk of FMCG Strategist 🎉 Retail segmentation has changed in India over time - but are we still using the early 2000 playbook in 2026 - GT vs. MT vs. Ecomm?? And we are seeing so many distribution strategies fail. Because the real segmentation today is not by labels like GT vs. MT vs. Ecomm? the real segmentation according to me is: "What behaviour does the outlet show consistently over time?" Here's the new reality.. I implemented this 1st hand in my commercial strategy roles and saw some significant shift in distribution performance: 1. Velocity Outlets - where products "fly" off the shelves **Velocity>format 2. Mission led outlets - Top up stores vs. monthly stock up stores, emergency stores vs. discovery stores, Convenience-first where speed matters more than choice.. **Mission decides assortment 3. Premium / affinity stores - Naturally pulls premium **Premium is the retail behaviour, not a consumer segment here 4. Discount led stores - These stores shape the grammage strategy 5. Influence led stores - Outlets that shape perception, not volumes (boutique, instagrammable ambience) **These outlets sell discovery Retail segmentation is no longer static. It's situational. A kirana near a school behaves differently from a kirana near a corporate park. A supermarket in Indore behaves differently from a supermarket in Mumbai. A q‑commerce dark store behaves differently based on neighbourhood income. Coming soon a "Retail Cheat-sheet 2026"; stay tuned for the next video on my YT channel @fromthedeskofFMCGStrategist as I’m building a list of behavioural indicators that teams can track to design retail strategies grounded in reality — not legacy labels. ✨I am also launching #weekend_insights_workshop — a 1‑hour, power‑packed virtual session on FMCG strategy and commercial insights. Limited seats only! DM me if you want to be on the guest list (nominal cost applicable) #FromthedeskofFMCGStrategist #Strategy #BusinessConsulting #Insights

40

2 Comments

📊 Mean vs Median vs Mode - Very Very Important 🤪 ❓ Which one should you use — and when? Choosing the wrong statistical measure can completely mislead your analysis. Here’s a simple rule-based breakdown 👇 🔢 Numerical Columns Mean → When data is symmetric & no outliers Median ✅ → Best when outliers exist Mode → When repetition matters 👉 Example: Salary, house prices → Median 🏷️ Categorical Columns Mean ❌ Median ❌ Mode ✅ 👉 Example: City, product, gender → Mode ⏰ DateTime Columns Median → Typical event time (robust) Mean → Average duration / interval 📝 Text Columns Mean ❌ Median ❌ Mode ✅ (most frequent word/value) 🔥 Golden Rule (Must Remember): If outliers exist → Avoid Mean 📌 Save this post for interviews, projects, and real-world analytics. 💬 Comment “YES ” If you agree 👍 #DataAnalysis #DataScience #Statistics #MeanMedianMode #Analytics #Python #SQL #JroshanCode #TechJroshan #InterviewPrep #Learning

13

2 Comments

🚀 Turning raw data into clear insights — watch it in action! 🎥 Using the UCI Online Retail dataset, an automated report generator cleans the data, summarizes key info, and creates insightful charts. Check out the video to see how the report comes to life! CC: Monal kumar Krish Naik KRISHAI Technologies Private Limited Sunny SavitaMayank Aggarwal #DataScience #Automation #Python #DataVisualization

14

Why Dashboards Often Fail Decision-Making 🚫📊 Too many dashboards show data… but not answers. Common pitfalls I see in MIS & Sales Operations dashboards: • Information Overload – too much data, no clarity • Data Confusion – metrics without context • No Actionable Insights – what should I actually do? • Misaligned Metrics – tracking what’s easy, not what matters 💡 The fix? Dashboards should guide decisions, not just display numbers. Highlight trends, focus on insights, and make actions obvious. Turn data into decisions — that’s where the real value is. #DataDriven #MIS #Analytics #DashboardDesign #SalesOperations #BusinessIntelligence #PowerBI

22