Most entry-level data science in Nepal is checking a table, cleaning it and explaining one number to a manager. Machine learning comes later, so the best first step is one small, honest analysis, and you can do it tonight in a free notebook with real Nepal data.
Below, step by step, you will load World Bank figures on electricity access, keep only Nepal, run the four checks analysts run before any chart, and draw one line chart. It takes about 30 minutes. You need a Google account and a browser.
What does data science in Nepal look like at entry level?
Job guides for Nepal name the same tools again and again. A January 2026 KumariJob guide to data science jobs in Nepal says analysts need "a proper understanding of tools like Excel, SQL, and Tableau". It names banking, telecoms, IT companies, e-commerce and healthcare as the sectors that hire.
Notice what is not on that list: deep learning. A junior analyst spends most of the week on tables. They find missing rows, fix wrong types, join two sources and write the result in plain words. A data scientist in Nepal starts from the same work, so it is the right thing to practise first.
Step 1: open a free notebook
Go to Google Colab and choose New notebook. Colab is a hosted Jupyter notebook service. Google's own FAQ says it is free of charge and "requires no setup to use".
The code runs on Google's computers, not on yours. So an old laptop is enough for this exercise, and nothing needs to be installed. Your notebooks are saved in your Google Drive.
A notebook is made of cells. You type code in a cell and press Shift+Enter to run it. Put each line below in its own cell.
Step 2: load the data
The data is "Share of the population with access to electricity" from Our World in Data. They take it from the World Bank and share it under the CC BY 4.0 licence, so you can use it in a portfolio as long as you name the source.
First cell: import pandas as pd
Second cell, the address of the file: url = "https://ourworldindata.org/grapher/share-of-the-population-with-access-to-electricity.csv?v=1&csvType=full&useColumnShortNames=true"
Third cell: df = pd.read_csv(url, storage_options={"User-Agent": "Our World In Data data fetch/1.0"})
The storage_options part sends a short name along with the request. The Python example on the Our World in Data chart page sends the same line, so copy it as it is.
Fourth cell: df.head(). You should see four columns: entity, code, year and eg_elc_accs_zs. The last one is the percentage of people with electricity. Short, odd column names are normal in real work. Find out what each column means before you use it.
Step 3: keep only Nepal
Fifth cell: nepal = df[df["entity"] == "Nepal"]
This keeps only the rows where the country is Nepal. Now run nepal.tail() to see the newest years. The last row should be 2024, with a value of 97.9.
Step 4: run four checks before any chart
This is the step most beginners skip, and it is the step a reviewer notices first. Run each check in its own cell and write the answer down.
-
How many rows?
nepal.shapegives the number of rows and columns. -
Any blanks?
nepal.isna().sum()counts missing values in each column. -
Any repeats?
nepal.duplicated().sum()counts rows that appear twice. -
Any missing years?
nepal["year"].max() - nepal["year"].min() + 1gives the number of years the data should cover. Compare it with your row count.
The last check is the one people miss. A year with no data usually does not show up as a blank. The row is just not there, so isna() says zero and the data looks complete. If the two numbers differ, some years are missing, and your chart will draw a straight line across the gap without telling you.
Then run nepal.describe() to see the smallest, largest and average values. The column is a percentage, so any value below 0 or above 100 is an error you must explain.
Step 5: draw one chart and write one sentence
Last cell: nepal.plot(x="year", y="eg_elc_accs_zs", title="Nepal: share of people with electricity")
Now write one sentence under the chart that a manager could read in ten seconds. Say what the number is, for which year and from which source. For example: "By 2024, 97.9 percent of people in Nepal had access to electricity, according to World Bank data via Our World in Data." Then add one line on what your four checks found.
That sentence is the real output of most analyst work. The chart only supports it. Save the notebook, add a short note on each check, and you have the start of a portfolio piece.
How to become a data scientist in Nepal from here
Our advice is plain. Do not start with a machine learning course. Learn pandas and SQL until you can run these four checks on any table without looking them up, and only then train a model. A model built on a table nobody checked gives confident wrong answers.
Next, try the same steps on a messier file, such as a sales sheet from a family shop or a college result sheet. Real files have blanks, repeated rows and dates typed three different ways. Fixing those is most of data analytics in Nepal.
Our Data Science and Machine Learning course in Nepal at Saarathi Academy in Old Baneshwor, Kathmandu, follows this order. Python gets one and a half weeks on its own before any library. Pandas comes in Week 3 and SQL in Week 4. Week 5 starts with a written data-check list run before any analysis: missing values, duplicates, wrong types and a row count that does not match the source. Machine learning starts in Week 9 of 12.
To compare courses, read our guide to choosing data science training in Kathmandu, which lists what a full curriculum should cover. Our week-by-week data science syllabus is public, so you can check it before you decide.
Common questions
What data science jobs in Nepal can a beginner aim for?
Most entry roles in data science in Nepal are data analyst, Power BI analyst and business analyst jobs. Banks, telecoms, IT companies and e-commerce firms post them. Pure research data scientist jobs are rare here, so many people start as analysts and grow from there.
Do I need a powerful laptop for data analytics in Nepal?
Not to start. Google Colab runs your code on Google's computers, so any laptop with a browser works for exercises like this one. Our course asks for at least 8 GB of RAM, and Power BI Desktop runs only on Windows.
What is the difference between data science and data analytics?
Data analytics answers questions about what already happened, using cleaning, SQL, charts and simple statistics. Data science adds prediction, such as a sales forecast or a model that flags customers likely to leave. Both start from the same checked and cleaned table.
How long does data science training in Nepal take?
It depends on the course, so always ask for a week-by-week syllabus. Our course runs 12 weeks, two hours a day from Monday to Friday, with about two hours of practice at home each day. It ends with a two-hour certification exam taken online at home.
Top comments (0)