Skip to content

LimaRods/Livepeer-Customer-Usage

Repository files navigation

Livepeer-Customer-Usage

My Take Home for the recruitment process of Livepeer

1. Introduction

This is the take-home for the recruitment process of Livepeer. The main purpose of this dashboard is to show my skills in Python, and analytics and answer the following questions given the daily customer usage data:

  • What are some high-level KPIs that you would recommend?
  • Who are our most valuable customers and why?
  • What customers have the most future potential and how did you target them?

2. Methodologies

2.1 KPIs

I used a personal framework to convert stakeholders' needs and statements into metrics. I set up 3 global metrics: Lenght of Transcoded Video per Period, Length of Video Streamed per Period, and Length of Video Uploaded per Period.

screenshot

It's possible to see the data from multiple angles such as total runnings, accumulated value, percent variation, histogram, and customer rankings. There are two types of ranking charts: an overall ranking considering all customers and another considering customer who created an account within the last 3 months from the end date of the filter.

2.2 Filters

The filters available are date range, date granularity, and ranking filter. The date range controls all dashboard elements, the date granularity change all charts and the ranking filter only affects the top customer's chart. In this way, it allows us to look at top customers within a date range.

screenshot

3. Data Analysis

You can switch among the metrics using tabs. Each tab view contains 6 charts, resulting in 18 total charts plus big numbers.

screenshot

4. Conclusions

  • Looking at the histogram for the transcoded duration, we can notice that few daily usages account for most of the total duration, and hence, few customers generate a relevant part of the total revenue from the streaming service. Surprisingly, the top 5 customers for transcoded duration account together for 90% of total transcoded videos.

screenshot

  • Between 1st Jan 22 and 21th Oct 22, the top one customer when it comes to transcoded duration is 5ef4. The top one new customer is a0c2 and represents almost 10% of the top 10th customer in terms of duration. We assumed the top new customers as future potentials.

screenshot

For video uploaded, we have 3fcf and 5b59 as the top customer and top new customer respectively

screenshot

About

My Take Home for the recruitment process of Livepeer

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published