Power BI research: what affects IT salary in Ukraine

What if we tell you there are several factors that can affect your salary in IT?

Knowing these factors, you can upgrade your skills according to the current market situation, and apply for a job you want. We wondered what these factors may be.

Look at what we found out!

client’s goals

The salary is one of the most important features of any job offer. A lot of candidates choose their jobs because of the salary, so our goal was to check what qualifications are highly valued among IT candidates.

During our BI research, we set two main hypotheses:

  • The first one was that a specialist’s gender may affect his or her salary. According to this hypothesis, male IT pros can get a higher salary than women in a similar position.
  • The second one was about the professional’s English level: the higher a person’s knowledge, the higher their salary will be.

our solutions

Data preparing

At this stage, we needed to prepare the data by:

  • removing physically impossible values
  • checking typos
  • checking missed values

We cleaned the information automatically using Pandas. These manipulations were required as a lot of respondents were too creative in their responses. 🙂

Data research

Machine learning data analysis stage helped us to understand how variables interact with each other, and to evaluate data distribution with descriptive statistics, visual methods and simple modeling.

Data modeling

At this stage, we used Machine Learning modeling (Keras and TensorFlow) to do our modeling and search for new hypotheses, if there any.

Sloboda studio used clusters, factor analysis methods, regression methods, and decided to use the decision tree. Our developers built a neural network with a three-layer architecture with 26 inputs and 1 output, which helped to find the variables that most affect the salary. So, at this stage, we were able to see if our initial hypotheses were right or wrong.

project stages

1. Initial stage

Data preparation

2. Modeling stage

Modeling, verification, and testing

3. Visualization

Creating visual data for the research

4. Deployment

Web deployment of all the data

TECHNOLOGIES

Tools (hosting, monitoring etc.)
Power BI
Integrations
Keras
TensorFlow

Timeline

December 2018

Team

1 machine learning developer
Project Results

For our IT services industry report, we got the following results:

  • The first hypothesis about a worker’s gender was rejected. According to our survey, this factor does not affect the specialist’s rate.
  • The second hypothesis told us about a person’s English level and the effect it has on salary. And we had a confirmation: really, the higher the knowledge, the bigger rate.
  • We found out that the key factor to affect a salary is a candidate’s experience. Therefore, a person’s qualifications and expertise are the most valuable factors for the market.

According to our research, the highest salaries can be found in Kyiv, Kharkiv, Odesa, and Lviv. The highest paid spheres are Gamedev, System Programming, iOS, and Multimedia. The highest paid positions are System Architect, Technical Lead, and Software Engineer.

Such statistical analysis methods in research are suitable for any subject area where deep understanding, analysis, and modeling are important.

Salary statistics is an excellent area for modeling, classifying, machine learning and statistical data analysis.

 

Contact the team, start like a rocket!