Integrated Reporting & Forecasting System for Outsourcing Firms

Introduction

The Reports & Forecast System is designed to deliver comprehensive reporting and forecasting capabilities on key performance indicators (KPIs) critical for outsourcing businesses. These indicators are calculated using real-time operational data, sourced from various existing systems within the company.

Challenge

Previously, the company lacked a unified platform for aggregating data across different departments. This fragmentation made it difficult to forecast developers’ workloads and company revenue accurately. The sales department struggled to determine how many deals needed to be closed to adequately utilise resources, while project managers faced challenges in forecasting overall resource allocation and project delivery timelines.

Solution

The implemented solution involved creating an integrated system that consolidates information from all departments and applications used within the company. Key functionalities include workload forecasting and budget estimation.

Integrated Systems & Utilised Data

BambooHR

Employee information:

• Department/Sub-department affiliation
• Hire and termination dates (where applicable)
• Details of absences, sick leave, and vacations

Atlassian Jira

Employee time tracking information:

• Amount of time spent
• Specific projects on which time was spent
• Types of recorded time: Billable to the client, Non-billable, Internal company work
  
  

Pipedrive

Sales pipeline data:

• Planned sales
• Position of potential projects in the sales funnel (probability of project initiation)
• Project start dates
• Planned workload based on project estimation (Google Sheets)

Smart Recruiter

Vacancy information:

• Vacancy status
• Predicted vacancy closure date

Internal Salaries Database

Employee salary details:

• Payment amounts
• Scheduled salary increases (if applicable)

Internal Employees Booking System

Resource allocation data:

• Planned or actual project start dates
• Planned or actual project completion dates

Internal Income System

Client invoicing details:

• Projects for which invoices are issued
• Payment status
• Payment dates or expected payment dates

Types of Reports & Forecasts

All reports or forecasts could be generated for specified time periods, representing historical facts if the period was in the past, forecasts if the period was future, or a combination if the selected period spanned both past and future.

• Company size report/forecast (Grouped by department or sub-department)
• Staff turnover report/forecast (Grouped by department or sub-department)
• Absence, sick leave, and vacation report/forecast (Grouped by department, sub-department, individual employees, or type of absence)
• Project count report/forecast (Grouped by department or sub-department)
• Sales report/forecast (Grouped by department or sub-department)
• Recruitment report/forecast (Grouped by department or sub-department)
• Payment reports (Grouped by project, client, department, or sub-department)
• Cash flow gap report (Grouped by project or client)
• Bench (resource availability) report/forecast (Grouped by department or sub-department)
• Salary budget report/forecast (Grouped by department or sub-department)
• Time expenditure report/forecast (Grouped by department, sub-department, type of work, or client billability)
• Revenue report/forecastа (Grouped by project, client, department, or sub-department)
• Overtime report (Grouped by project, client, department, or sub-department)

Anomaly Detection & Notifications

Bidirectional LSTM Autoencoder for detecting anomalies in both historical and forecasted data.

Periodic checks for anomalies, alerting responsible users about:

• Budget overruns
• Spikes in employee absences
• Employee downtime
• Increases in cash flow gaps
• Underreported employee hours
• Employee overtime

Technologies

Programming Language:

• Python

Libraries:

• sentry
• pytest
• requests
• jsonpickle
• raven
• psutil
• loguru
• dotenv
• google-api-python-client
• google-auth
• MinMaxScaler

Linters

• black
• mypy
• flake8
• ruff

Databases

• MySQL
• SQLite

APIs Utilised:

• BambooHR
• Google Sheets
• Smart Recruiter
• Jira
• Pipedrive
• 3 internal APIs
• Slack

Neural Networks (AI/ML):

• TensorFlow
• Bidirectional LSTM Autoencoder

Conclusion

The implemented Reports & Forecast System effectively integrates data from various departments and company applications, creating a centralised information hub. It addresses significant challenges related to employee workload visibility, revenue forecasting accuracy, and inter-departmental planning complexities. By connecting with systems like BambooHR, Jira, Pipedrive, and others, the system enables detailed monitoring and analysis of employee, project, sales, and financial data. Consequently, this leads to more precise planning, improved resource management, and accurate forecasting of key business metrics. Additionally, the integration of a trained Bidirectional LSTM model enhances anomaly detection capabilities, contributing further to informed decision-making and improved operational efficiency.