On-Call Ticket Volume Forecasting

SupportDesk Inc.'s manager must forecast next month's ticket volume from sparse, noisy historical data — manually, before any AI-assisted forecasting is allowed. A mid-dataset feature launch has muddied the trend.

# Aplly.xyz Case Study Submission

## Title
On-Call Ticket Volume Forecasting

## Type
Data Analytics

## Difficulty
Beginner

## Estimated Time
45 minutes

## Overview
SupportDesk Inc.'s manager must forecast next month's ticket volume from sparse, noisy historical data — manually, before any AI-assisted forecasting is allowed. A mid-dataset feature launch has muddied the trend.

## Case Details

Function Focus: Trend estimation, manual averaging/weighting, outlier judgment

Scenario:
A new feature launched mid-year caused a ticket spike. The manager must decide how much of that spike is "new normal" versus temporary noise, and produce a defensible forecast for next month without automated trendlines.

Dataset Structure:
- Month, Tickets Opened, Tickets Closed, Feature Launch (Yes/No), Team Headcount

Tasks:
1. Manually tabulate or hand-chart the trend (pen/paper or plain grid — no spreadsheet formulas)
2. Decide which data points are outliers and justify why in writing
3. Produce a single forecasted number for next month with your reasoning shown step by step
4. Only after submitting: run the same data through any tool/AI forecast and compare deltas, noting where your manual reasoning diverged

Expected Output:
Forecast number + written reasoning trace + a short post-hoc comparison note against the tool-assisted forecast.

Evaluation Criteria:
Reasoning transparency, quality of outlier judgment, and thoughtfulness of the gap analysis between manual and tool-assisted forecasts.

## Data Sources

| Month | Tickets Opened | Tickets Closed | Feature Launch | Headcount |
|---|---|---|---|---|
| Jan | 420 | 410 | No | 8 |
| Feb | 445 | 430 | No | 8 |
| Mar | 610 | 500 | Yes | 8 |
| Apr | 580 | 570 | No | 9 |
| May | 460 | 455 | No | 9 |
| Jun | 500 | 490 | No | 9 |
| Jul | 515 | 505 | No | 9 |
| Aug | 495 | 490 | No | 9 |
| Sep | 530 | 520 | No | 10 |
| Oct | 540 | 535 | No | 10 |

## Solution Frameworks
Manual trend/outlier analysis, weighted recency averaging

## Solver Guidance & Tutorials
Link to: "Reading Trends by Hand Before Trusting a Model" tutorial

## What You'll Learn
- Estimation without automation
- Distinguishing signal from noise in operational data
- Self-auditing manual conclusions against AI output rather than deferring to it

## Tags
forecasting, support operations, data judgment, trend analysis

## Registration Links
- Register as Solver
- Register as Evaluator

| Month | Tickets Opened | Tickets Closed | Feature Launch | Headcount |
|---|---|---|---|---|
| Jan | 420 | 410 | No | 8 |
| Feb | 445 | 430 | No | 8 |
| Mar | 610 | 500 | Yes | 8 |
| Apr | 580 | 570 | No | 9 |
| May | 460 | 455 | No | 9 |
| Jun | 500 | 490 | No | 9 |
| Jul | 515 | 505 | No | 9 |
| Aug | 495 | 490 | No | 9 |
| Sep | 530 | 520 | No | 10 |
| Oct | 540 | 535 | No | 10 |

Manual trend/outlier analysis, weighted recency averaging