Python and Pandas for Data Manipulation

Many analytics teams still rely on spreadsheets for work that breaks down at scale: inconsistent cleanup steps, manual merges, and reports that are hard to reproduce. Python and Pandas for Data Manipulation gives IT Directors and Training Managers a standard, programmatic workflow their teams can share.

Role fit: This is intermediate, associate-level training for data analysts and IT practitioners who support reporting. The trainer reviews basic Python data types, but notes the course is harder without Python fundamentals. For learners new to Python, the CBT Nuggets catalog includes Introductory Python for Data Analysts as a foundational option.

Time investment: Plan for about 18.5 hours across 20 skills, in sessions of roughly 45–70 minutes. Learners can complete one skill per sitting without pulling them off project work for long stretches.

Change management: For staff coming from Excel, the course addresses the shift head-on. It sets up a repeatable Anaconda, conda environment, and Jupyter Notebook startup and shutdown routine that becomes habit over time. It then builds from Series and DataFrames to cleaning, filtering, joins, time series data, and visualization.

How it supports the outcome: Starter and completed notebook files, challenge questions, and open-ended exercises on real CSV datasets reinforce each skill. Add the course to a CBT Nuggets Playlist to sequence it after foundational Python training and onboard new analysts consistently.

Data rarely arrives clean, in one file, or in one format. This course trains your team to handle that reality programmatically in Pandas and Jupyter Notebook, an alternative to manual spreadsheet work for analysts coming from Excel.

The scenarios mirror everyday analyst work. Learners work with CSV and Excel files, including modified Kaggle datasets, then clean, filter, sort, and deduplicate them. They combine data from multiple sources in reusable notebook workflows. The course ends with a cryptocurrency portfolio built from data pulled from the internet with DataReader, with no web scraping required, then visualized in the notebook.

• Faster, repeatable data prep: Analysts handle null values, optimize data types, sort, and filter by condition in reusable notebook code. This replaces one-off manual edits.
• Reliable multi-source reporting: Team members combine related tables so reports draw on connected data instead of separate spreadsheets.
• Cleaner text and grouped summaries: Learners split and strip messy text columns, then use GroupBy operations to summarize data by category, such as sector.
• Time-aware analysis and visuals: Practitioners work with timestamps, DatetimeIndex, and date ranges, then build charts and visual summaries for stakeholders.

The open-ended challenges, such as finding a CSV and performing operations on it, build the independent problem-solving that Team Leads need before handing analysts real datasets.

Knowledge

• How pandas, Jupyter Notebook, Anaconda, and conda fit together, including conda environments and the startup and shutdown routine for each work session
• How the pandas Series and DataFrame data structures work, plus the attributes and methods used to inspect them and pull metadata from them
• The main approaches to handling null values, filtering duplicates, and changing data types, such as categories, to make datasets smaller and faster to work with
• How GroupBy objects differ from DataFrames, and when to use concatenation, inner joins, outer joins, or full outer joins to combine data
• How Python datetime objects relate to pandas Timestamp, DatetimeIndex, and the date_range method for time series work
• How loc and iloc access data by label or position when extracting information from pandas objects

Ability

• Set up a working Python and pandas environment, then run code cells in Jupyter Notebook using keyboard shortcuts
• Import external CSV files, including course datasets such as student performance data, then clean, sort, filter, and deduplicate them
• Manipulate text columns by splitting and stripping values, for example separating combined names into their own columns
• Group records by a field such as sector and aggregate the results with operations like sum and mean
• Merge related tables with pandas joins and concatenation to create a single view
• Pull time series data with DataReader without writing web scrapers, then build and style Matplotlib visualizations