Active Project Work: Data Collection, Cleaning, Analysis, and Visualization

What This Phase of the Externship Requires

You have moved from planning into doing. The Learning Agreement you signed defines the scope of your work—specific datasets, tools, deliverables, and timelines. During this active project phase you will collect or ingest real data, clean it, analyze it, and produce visual outputs that a supervisor will formally review. Everything you produce must meet professional quality standards, not just class standards. That means accurate results, clear documentation, and responsible handling of any sensitive data.

Daily Work Rhythm

Treat every worksite day like a professional workday. Arrive on time, clarify your task list at the start of the shift, and close out your work session by committing all code and documentation to the employer's version-control system or shared workspace. A commit that never arrives is work that cannot be reviewed, audited, or built upon by colleagues.

Typical daily tasks may include writing SQL queries against production or staging databases, cleaning CSV or API-sourced files in Python or R, updating Power BI or Tableau dashboards, or running descriptive and predictive analyses. Regardless of the task, apply the same habits every day: comment your code, name files clearly, and log what you did in your externship journal before you leave.

Data Privacy and Responsible Handling

Many worksite datasets contain personally identifiable information (PII) or proprietary business data. You are legally and ethically obligated to protect it.

• Work only on systems and files your supervisor has explicitly authorized.
• Do not copy production data to personal devices, personal cloud accounts, or any unapproved storage.
• Follow the employer's data-classification and retention policies exactly.
• If you accidentally see data you were not meant to see, notify your supervisor immediately rather than ignoring it.
• When in doubt about whether an action is permitted, stop and ask.

These are not suggestions—violating data-privacy rules can end a placement, expose the employer to legal liability, and follow you professionally. The Employer Quality Rubric scores adherence to data-privacy policy as one of its five required criteria.

Ingestion and Cleaning Best Practices

Raw data is almost never analysis-ready. Before you can trust any result, you must document and address data-quality issues. A professional cleaning workflow includes:

1. Profile the data first. Count rows, check column types, identify nulls, and look at value distributions before writing a single transformation.

2. Document every decision. If you drop rows with missing values, record why. If you impute, record the method and rationale.

3. Keep the raw file untouched. Work on a copy or use a pipeline that preserves the original; this supports reproducibility and audit.

4. Validate after each transformation. Row counts, summary statistics, and spot checks catch errors introduced by cleaning steps themselves.

5. Produce a data-quality report. Your mid-point deliverable may be exactly this: a cleaned dataset accompanied by a written summary of what was found and what was done about it.

Exploratory and Descriptive Analysis

Exploratory data analysis (EDA) is the process of summarizing the main characteristics of a dataset, usually through statistics and plots, before formal modeling. During EDA you are asking: What does this data actually contain? Are there outliers? What relationships exist between variables?

Record your EDA findings in your journal. Screenshots of distributions, correlation matrices, or pivot-table summaries count as evidence of analytical thinking. Supervisors and credential exam objectives both expect you to distinguish between descriptive statistics (what happened) and predictive or inferential work (what might happen or why).

Visualization Standards

A visualization that misleads is worse than no visualization at all. Follow these principles:

• Choose chart types that match the data: bar charts for categories, line charts for time series, scatter plots for relationships.
• Label axes clearly and include units.
• Do not truncate y-axes in ways that exaggerate differences.
• Use color purposefully—accessible palettes work for colorblind viewers.
• Write a title and, where appropriate, a one-sentence insight caption so the chart communicates without requiring a verbal explanation.

Power BI and Tableau both support these practices. Your dashboard or chart must be readable by a business stakeholder who is not a data specialist.

Team Meetings and Status Updates

You are required to attend and contribute to at least four team meetings, stand-ups, or sprint reviews. Before each one, write a brief status update: what you completed, what is in progress, and any blockers. At each meeting, request feedback on at least one piece of work. Record both the feedback and any action items in your journal. This habit builds the professional communication skills that every credential exam and every employer expects.

The Mid-Point Review

At approximately the 160-hour mark, you will present one completed portfolio artifact to your site supervisor for a formal mid-point review. The artifact might be a cleaned dataset with a data-quality report, a functional dashboard with a written user guide, or an analysis memo with visualizations and recommendations.

Your supervisor will score it using the program's Employer Quality Rubric across five criteria: accuracy, documentation, visualization clarity, professional communication, and adherence to data-privacy policy. A passing score is 3 out of 5 or higher on every criterion—not an average, but every single criterion. If any criterion scores below 3, you must revise and resubmit within one week.

Using Credential Checklists as Self-Assessment Tools

Keep the CompTIA Data+ exam objectives, the Google Data Analytics competency checklist, and the IBM Data Science Professional checklist at your workstation (digital or printed). At the end of each week, open your journal and mark which competencies your day-to-day work touched. Gaps you notice now are gaps you can fill before the credential exams, not after.