Dataford
Interview QuestionsInterview GuidesExperiencesMock InterviewsPricing
Get started

Stock Price Forecasting Approach

HardMachine Learning00:00
Practice interviewer
In session
5 left
00:00

Your question is Stock Price Forecasting Approach. Take a moment with it on the right.

Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).

You need to log in / sign up to chat or submit.

Problem

Scenario

You are asked to build a model that predicts stock prices from historical market data. The goal is to produce forecasts that could support downstream investment research, while avoiding common mistakes like leakage and unrealistic backtests.

Question

How would you approach building a predictive model for stock prices?

Representative Dataset

size·8 years of daily data for 520 equities, about 1.05M ticker-day rowstarget·Next-day log return, plus optional next-day closefeatures·OHLCV, lagged returns, rolling stats, momentum, sector and macro contextmissing_data·Sparse market gaps and release-driven macro missingness