Analyze performance of system operations
Budget: $15 – $25 USD
This data analysis project is meant to help us understand performance of certain operations in our system. Each operation is represented by a CSV record containing certain attributes of the operation and the execution time.
The data is represented by a set of approximately 30K records in CSV format (in two files), covering the time period between June 15 and September 20 of 2021 and between May 6 and September 20 of 2022 (with an acknowledged gap between September 20, 2021 and May 6, 2022).
We’d like for that data to be analyzed and graphed to help answer the following questions:
1. What is the overall trendline graph for performance, with a sliding ‘window’ of about 2 weeks in duration? (So it’s not just a straight line, but is rather a multi-segment line of different slopes, allowing us to discern possible significant changes in averages.)
2. What are the approximate dates (or short periods) when there seemed to be a noticeable and persistent performance change?
3. A table of mean and median performance numbers by the following operation attributes:
a. Program
b. Measure
c. Measure type
4. Once we understand the results from item (3), we will likely want to answer questions (1) and (2) for ‘top 5’ slow performing programs, measures, and measure types, to attempt to understand the root causes of performance deterioration for each of them.
If possible, we would like to get the ‘answers’ as Google Sheets documents, in case we wanted to do some further analysis ourselves in the future. On the other hand, if you feel like you can do this more quickly and efficiently in a tool like R-Studio, and just provide us with the numbers and the ‘answers’, that’s fine as well.
We will provide the data files once the project has been accepted.
The data is represented by a set of approximately 30K records in CSV format (in two files), covering the time period between June 15 and September 20 of 2021 and between May 6 and September 20 of 2022 (with an acknowledged gap between September 20, 2021 and May 6, 2022).
We’d like for that data to be analyzed and graphed to help answer the following questions:
1. What is the overall trendline graph for performance, with a sliding ‘window’ of about 2 weeks in duration? (So it’s not just a straight line, but is rather a multi-segment line of different slopes, allowing us to discern possible significant changes in averages.)
2. What are the approximate dates (or short periods) when there seemed to be a noticeable and persistent performance change?
3. A table of mean and median performance numbers by the following operation attributes:
a. Program
b. Measure
c. Measure type
4. Once we understand the results from item (3), we will likely want to answer questions (1) and (2) for ‘top 5’ slow performing programs, measures, and measure types, to attempt to understand the root causes of performance deterioration for each of them.
If possible, we would like to get the ‘answers’ as Google Sheets documents, in case we wanted to do some further analysis ourselves in the future. On the other hand, if you feel like you can do this more quickly and efficiently in a tool like R-Studio, and just provide us with the numbers and the ‘answers’, that’s fine as well.
We will provide the data files once the project has been accepted.