Data Analysis Interview Questions

Katherine B. - "Before jumping into solutions, I’d want to first understand the feature launch’s objectives: what metrics were we trying to move and how did those top-line metrics perform. In an ideal scenario, we’d A/B test the feature so we can statistically measure metric performance over the existing features. If we weren’t able to A/B test, we would evaluate performance using pre-post on our key metrics we’re aiming to move, as well monitor funnel performance. Likewise, I’d make sure that engagement, ado"See full answer

Saurabh K. - "Goals : Determine if the TV series should be renewed If it should be renewed, how much should Netflix be willing to pay for this series Let's assume that the goal is to maximize subscriber retention and engagement while paying a reasonable amount for the licensing costs that is justified by the value added by the series. Assumptions : The show is exclusive to Netflix for a particular region (for eg. US) It has been on the platform for an year Netflix has subscriber level data around"See full answer

Himanshu G. - "To model ROI for a product launch, the first step is to define the timeline you're targeting Example 6 months post-launch, 1 year, or even 5 years. Tip: Start with a 1-year ROI projection to estimate near-term returns, and build a 3-year projection to evaluate growth and scalability. ROI is essentially the net return over that period: Profit=Revenue (within timeline)−Total Cost (from project start) Total Cost includes both fixed and variable costs incurred since t"See full answer

Sergio G. - " import pandas as pd def findunsoldproducts(transactions: pd.DataFrame, products: pd.DataFrame) -> pd.DataFrame: productsnotsold = products[~products["id"].isin(transactions["product_id"].unique())] productsnotsold = productsnotsold[["id", "name", "stock"]] productsnotsold = productsnotsold.sort_values("id") return productsnotsold debug your code below products = pd.DataFrame({ 'id': [1, 2, 3, 4, 5, 6], 'name': ['ePhone 10', 'Instant Cooker 200"See full answer

Catherine T. - "We want sales to grow, in order to have a growth in revenue. And customer usage as well as it allows to see if our product lead more engagement from our users. So to be able to see this overall evolution I would make a line chart for both : Sales : with month on x-axis and sales revenue on y-axis Customer Usage : with month on x-axis and a KPI allowing to measure customer usage (nblogins or nbsessions or nbgamesplayed, ... depending on the industry) on y-axis Moreover, after knowing th"See full answer

Kusheta K. - "“I once worked with a dataset that had missing and inconsistent tracking. I first evaluated data quality, cleaned what was reliable, and documented assumptions. Then I collaborated with engineers to implement improved event tracking. This ensured more accurate analysis in the future and improved decision-making.”"See full answer

Kevin T. - "When a stakeholder’s request is ambiguous, I start by clarifying the goal and defining what “success” looks like. I ask targeted questions to understand the business problem, the timeframe, the scope/location, and who the analysis is for. Then I confirm definitions (metrics, segments, filters), agree on the expected output format, and restate the request back to them in one sentence before I begin."See full answer

Arshad P. - "Schema is wrong - id from product is mapped to id from transactions, id from product should point to product_id in transcations table"See full answer

Gabriel P. - "Hi, my solution gives the exact numerical values as the proposed solution, but it doesn't pass the tests. Am I missing something, or is this a bug? def findrevenueby_city(transactions: pd.DataFrame, users: pd.DataFrame, exchange_rate: pd.DataFrame) -> pd.DataFrame: gets user city for each user id userids = users[['id', 'usercity']] and merge on transactions transactions = transactions.merge(user_ids, how='left"See full answer

Katherine B. - "We want to use rigorous framework for evaluating shipping a new feature — ideally an A/B test. If an A/B test is not available, we first evaluate quantitative data; we look at feature adoption metrics, time-to-use, retention and frequency of visitation. What does the business impact of the feature on conversion rates, revenue per users and LTV, and secondarily evaluate any error rates that could be occurring after the launch of the new feature. It’s important for this analysis to perform segmen"See full answer