OnboardIQ – Employee Onboarding Analytics Dashboard

Problem Statement

A rapidly scaling startup stores employee onboarding progress, internal tool usage, and support request history separately, making it difficult to identify operational bottlenecks affecting employee productivity.

Project Objective

Build an interactive analytics dashboard that consolidates onboarding data, analyzes productivity trends, and provides actionable insights using Python, Pandas, NumPy, SQLite, and Streamlit.

Technology Stack

• Python
• Pandas
• NumPy
• SQLite
• Streamlit
• GitHub Actions
• Jupyter Notebook
• Git & GitHub
• Figma

Repository Structure

src/
data/
    raw/
    processed/
database/
notebooks/
reports/
tests/
docs/
design/

Data Sources

The project currently works with five datasets covering the major dimensions of the onboarding process:

1. Employee Dataset

   • Employee master/reference information
   • Used as the primary employee dimension

2. Onboarding Progress Dataset

   • Onboarding tasks, stages, statuses, due dates, and completion progress

3. Internal Tool Usage Dataset

   • Aggregate employee activity across approved internal productivity tools

4. Support Ticket Dataset

   • Structured support requests, issue categories, priorities, statuses, and resolution times

5. Learning Management Dataset

   • Training assignments, completion status, assessment scores, and completion time

All datasets are currently stored in the raw data layer and are subject to validation and processing before analytical use.

Data Validation & Profiling

Dataset intake, validation, and profiling have been completed for all five source datasets.

The profiling workflow includes:

• Dataset shape and schema inspection
• Column and data-type validation
• Primary-key uniqueness checks
• Null primary-key checks
• Missing-value analysis
• Referential-integrity validation
• Date-order validation
• Numeric range validation
• Categorical-value profiling
• Status and business-rule validation
• Data-quality findings documentation

Current Profiling Results

Dataset	Rows	Columns	Primary Key	Duplicate IDs	Null IDs	Referential Integrity
Employee	311	36	EmpID	0	0	N/A
Onboarding Progress	4,222	9	onboarding_task_id	0	0	Pass
Internal Tool Usage	13,995	9	usage_id	0	0	Pass
Support Tickets	1,048	8	ticket_id	0	0	Pass
Learning Management	3,192	9	learning_record_id	0	0	Pass

Date Validation Findings

• No employee records were found where termination occurred before the hire date.
• No learning records were found where course completion occurred before assignment.
• 1,788 onboarding tasks were completed before their due dates.
• 960 onboarding tasks were completed after their due dates.
• Late onboarding tasks will be retained as a potential indicator of onboarding delays and operational bottlenecks.

Data Quality Findings

The profiling stage identified data-quality considerations that will be addressed during processing and cleaning, including:

• Missing values in fields where nulls may represent legitimate business states
• Date columns currently stored as string/object types
• Categorical values requiring whitespace normalization
• Controlled-value standardization
• Business-rule validation for task, learning, and support statuses

Raw source datasets are not modified during the profiling stage.

Current Progress

Product & Design

• ✅ Product Requirements Document (PRD) completed.
• ✅ MVP scope finalized, including features and exclusions.
• ✅ Data requirements and source definitions documented.
• ✅ Data dictionary completed for all five datasets.
• ✅ Low-fidelity wireframes designed.
• ✅ Unified dashboard and supporting pages structured for the MVP.
• ✅ Settings page updated with Light and Dark theme options.
• ✅ Reports & Export page updated to support CSV and PDF exports.

Data Engineering & Analysis

• ✅ Five source datasets identified and added to the project.
• ✅ Raw data directory structure created.
• ✅ CSV data ingestion workflow established.
• ✅ Dataset loading and initial validation implemented.
• ✅ Dataset profiling notebook created.
• ✅ All five datasets profiled for data quality.
• ✅ Primary-key and duplicate validation completed.
• ✅ Referential-integrity checks completed.
• ✅ Missing-value analysis completed.
• ✅ Date and numeric validation completed.
• ✅ Categorical and status profiling completed.
• 🚧 Data cleaning and transformation in progress / next stage.
• ⏳ Processed analytical dataset creation pending.
• ⏳ SQLite database integration pending.
• ⏳ Analytical metrics and KPI development pending.

Dashboard & Application

• ⏳ Dashboard implementation pending.
• ⏳ Streamlit application development pending.
• ⏳ Interactive filters and visualizations pending.
• ⏳ Productivity and onboarding KPI implementation pending.
• ⏳ Final reports and export integration pending.

GitHub Development Workflow

The project follows a structured GitHub workflow:

Issue
  ↓
Branch
  ↓
Development
  ↓
Commit
  ↓
Pull Request
  ↓
Review
  ↓
Merge

Development work is organized through GitHub Issues, milestones, branches, commits, and pull requests.

Design

Low-Fidelity Wireframes

Figma Link:

https://www.figma.com/make/i6QNepl4k69jR7MHq3uifz/Data-Analytics-Dashboard---Low-Fid--Copy-?fullscreen=1&t=tpzJzT7vpdmrSJqB-1&code-node-id=0-6

Data Profiling Notebook

The initial data-quality profiling workflow is available at:

notebooks/
└── 01_data_source_profiling.ipynb

The notebook can be executed from a fresh kernel using Run All to reproduce the profiling results.

Run the App

The Streamlit application will be run using:

python app.py

Application implementation is currently pending.

Project Status

🚧 Data Preparation & Development Phase

Completed

• ✅ PRD Completed
• ✅ MVP Scope Defined
• ✅ Low-Fidelity Wireframes Completed
• ✅ Data Sources Finalized
• ✅ Data Dictionary Completed
• ✅ CSV/JSON Data Ingestion Workflow Established
• ✅ Dataset Profiling Completed
• ✅ Data Quality Assessment Completed
• ✅ GitHub Project Workflow Established

In Progress / Next

• 🚧 Data Cleaning & Transformation
• ⏳ Processed Dataset Creation
• ⏳ SQLite Database Setup
• ⏳ Analytical KPI Development
• ⏳ Streamlit Dashboard Development
• ⏳ Power BI-Compatible Data Exports
• ⏳ Testing & Validation
• ⏳ Final Dashboard Integration

Project Objective Outcome

The final OnboardIQ product will provide a consolidated view of employee onboarding health by combining:

Employee Data
      +
Onboarding Progress
      +
Internal Tool Usage
      +
Support Tickets
      +
Learning Management
      ↓
Data Cleaning & Transformation
      ↓
SQLite / Analytical Layer
      ↓
KPI & Metric Calculation
      ↓
Streamlit Dashboard
      ↓
Onboarding Insights & Bottleneck Identification

The intended outcome is to help stakeholders identify onboarding friction, monitor completion and productivity proxies, understand support and learning patterns, and identify operational factors that may delay new-hire productivity.