egytech_api Wrapper Showcase

Installs and imports

Install the latest versions of egytech_api and plotly

!pip install egytech_api==1.0.1 plotly==5.22.0

Import necessary functionality

from egytech_api.models import ParticipantsQueryParams from egytech_api.core import PoolingClient import plotly.express as px

Fetch Data From The egytech API

Construct The Queries As A List of ParticipantsQueryParams

queries = [ ParticipantsQueryParams(title="backend", cs_degree=True), ParticipantsQueryParams(title="frontend", cs_degree=True), ParticipantsQueryParams(title="fullstack", cs_degree=True) ]

Initialize The PoolingClient Instance

client = PoolingClient(queries=queries)

Generate And Inspect The Resulting pandas.DataFrame

df = client.get_df() df.head()

Creating Some Cool Visualizations

Scatter Plot

Here, we create a scatter plot to view salary increase with the increase in experience across three different job titles, namely, "backend", "frontend", and "fullstack".

fig = px.scatter(df, x="Yoe", y="TotalCompensationEgp", color="Title") fig.update_layout( title_text = "Salary Progression With Experience In Web Developers From CS Backgrounds" ) fig.update_yaxes(maxallowed=200000) fig.update_xaxes(maxallowed=15) fig.show()

Here, we notice that, among web developers from a computer science background, back-end engineers often get paid higher than the other two categories of web developers, while also being paid more in lower experience contexts. We also notice that egytech's data contain a few veteran full-stack engineers having >= 15 years of experience, yet being compensated poorly compared to those who are exclusively back-end engineers.

Box Plots

Here, we create a box plot to further inspect the compensation data by title:

fig = px.box(df, x="Title", y="TotalCompensationEgp", points="all") fig.show()

We can infer that on average, back-end engineers get paid higher than front-end and full-stack engineers in Egypt. However, there are many more back-end engineers in the data than there are of the other categories, so this could make the data in the smaller samples incorrectly skewed.