AI, machine learning and deep learning · Introductory

Look at the Big Picture

Written by SignalsLab

Early edition: these lessons are being reviewed.

Content

Summary

This lesson introduces a machine learning project designed to predict California district median housing prices using census data, aiming to replace a costly and error-prone manual estimation process. It emphasizes that defining the business objective is the critical first step, as it dictates the problem framing, algorithm selection, and performance measures. The goal is to create an automated model that feeds into a downstream investment decision system, improving accuracy and efficiency over the current expert-based approach.

Intuition

Think of machine learning not just as coding, but as solving a specific business problem. Before writing any code, you must understand why you are building the model. In this case, the business wants to invest in real estate but currently relies on human experts to estimate housing prices. This manual process is slow, expensive, and often wrong by more than 30%. By building an ML model, we aim to automate this estimation, providing faster and more accurate data to a downstream system that decides where to invest. The key insight is that the model is just a tool; the real goal is the business outcome—better investment decisions. Therefore, you must always start by understanding the current manual solution and the final business objective, rather than jumping straight into algorithm selection.

Derivation

  1. Identify the business objective: Replace manual housing price estimation with an automated model to support investment decisions.
  2. Analyze the current solution: Experts manually estimate prices using complex rules, which is costly, time-consuming, and has errors exceeding 30%.
  3. Define the data source: Use census data including population, median income, and median housing price for California districts (block groups with 600-3,000 people).
  4. Determine the model's role: The model predicts median housing price, which serves as input for a downstream ML system that determines investment viability.
  5. Establish the workflow: Start with a structured project checklist, adapting it to specific needs, to ensure all steps from problem framing to deployment are covered.

Worked example

Problem

A company currently uses experts to estimate housing prices for investment. The experts take 2 days per district and are wrong by 30% on average. You are asked to build an ML model. What is the primary business objective, and what is the baseline for comparison?

  1. Identify the business goal: The objective is to improve accuracy and efficiency of price estimation to feed into an investment decision system.
  2. Identify the baseline: The current manual estimation by experts is the baseline. It takes 2 days and has >30% error.
  3. Determine success criteria: The ML model must be faster than 2 days and have an error rate significantly lower than 30% to be considered successful.
  4. Select approach: Since the output feeds into a downstream system, the model's accuracy directly impacts revenue, making precise framing essential.

Answer The primary business objective is to automate and improve the accuracy of housing price estimation to support real estate investment decisions. The baseline for comparison is the current manual estimation process, which has an error rate exceeding 30% and takes 2 days per district. The ML model must outperform this baseline in both speed and accuracy.

Common mistakes

  • Assuming that building the model is the end goal rather than a means to achieve a business objective.
  • Ignoring the current manual solution, which provides a performance baseline and insights.
  • Treating the project checklist as rigid rather than adaptable to the specific project needs.

Animation

Animation coming soon

The animated version of this lesson is being prepared for the web. In the Signals AI app, lessons are drawn step by step.

Code

This is a text lesson: it has no animation code yet.

ccBYNC© 2026 SignalsLab. Lesson text shared as is under CC BY-NC 4.0.