Journal logo

How to Estimate a Data Analytics Project Timeline

A 4-step process to answer "how long will this take?"

By Mad For FabricPublished 3 years ago 3 min read
Like
How to Estimate a Data Analytics Project Timeline
Photo by Immo Wegmann on Unsplash

In your role as a data scientist or data analyst, a common question you’ll get from stakeholders is “how long will this take”? When I first become a data scientist, I found the question challenging to answer. If I said it took a month to build a model and I couldn’t deliver, would that harm my stakeholder’s perception of my abilities? If I built the model faster, would my stakeholder think I was prone to overestimating? If you find yourself struggling to provide better time estimates as I did, try this 4-step process I’ve developed that has helped me provide more accurate timelines.

1. Determine Project Scope

Set up a meeting with your stakeholders to define the project scope and expected deliverables. Sometimes stakeholders may not know exactly what they want. Clarify the business goal and questions stakeholders want to address in this project.

For example, marketing asks you to build a model to predict customer churn. The scope of the project can vary greatly depending on how the model will be used. If marketing plans to send emails to customers who are likely to churn, you’ll need more time after building the model to put the model into production if there’s no data engineer to help. If marketing just wants to know the top factors that predict a customer’s likelihood to churn, this will be an ad hoc model that doesn’t need to go into production.

2. Confirm Data Availability

After you determine the project scope, confirm that the data needed to complete the project is available. Also, check there’s enough historical data needed to build a meaningful model or run a detailed analysis.

If you need to bring new data into the data warehouse, budget extra time to do this. If you require data engineers or another team to help, make sure to caveat this as a dependency for project completion. This step is critical because if there are no resources to help you bring in the data needed, the project will need to be deferred.

3. List Project Tasks

Create a list of high-level tasks for the project and estimate the time needed to finish each. For example, to build the customer churn model an example of the tasks could be:

  • Exploratory data analysis
  • Data preparation and feature engineering
  • Model development
  • Model tuning
  • Summarize model results for stakeholder signoff
  • Productionize model

Add up the time for all tasks and you’ll have an initial estimate of the time needed to complete the project.

4. Apply a Fudge Factor

When you work on a project there are always unexpected events you didn’t budget time for — longer data preparation, model run time, or model tuning. You could need help from other teams and this adds to the project time. To account for these unforeseen cases, apply a fudge factor. The fudge factor should also account for the time you normally spend in meetings and answering questions that take away time you can spend on the project.

Use estimated and actual completion times for past projects to figure out your fudge factor. If you estimated 20 days but it actually took 30 then your fudge factor is 30 divided by 20 or 1.5. If you estimate the current project will take 25 days in step 3, take 25 times 1.5 and give 38 days ( or about 7.5 weeks ) as the estimate to complete the project.

Final Thoughts

As a data scientist or data analyst, getting asked “how long will this take” doesn’t have to be a difficult question to answer.

To recap, use this 4-step process:

  1. Determine the project scope.
  2. Confirm data availability needed for the project.
  3. Define high-level tasks and sum the estimated time for each task.
  4. Apply a fudge factor to the total from step 3 to account for unexpected events, time spent in meetings, and answering questions.

Thanks for reading and if you found this article helpful, please consider leaving a tip.

workflow
Like

About the Creator

Mad For Fabric

Sewist and fabric obsessed. Sharing my creative journey one story at a time. Blogging about my creations at www.madforfabric.com

Reader insights

Be the first to share your insights about this piece.

How does it work?

Add your insights

Comments

There are no comments for this story

Be the first to respond and start the conversation.

Sign in to comment

    Find us on social media

    Miscellaneous links

    • Explore
    • Contact
    • Privacy Policy
    • Terms of Use
    • Support

    © 2024 Creatd, Inc. All Rights Reserved.