How to Ace Your DATA4400 Forecasting Assessment 3

Comprehensive Guide

Quick Navigation

PHASE 1: PLANNING AND PREPARATION (Weeks 10-11)

▼

Step 1: Select Your Plan and Business Context

Choose the most feasible plan:

  • Plan A (Ideal): Use data from an organization you work/worked for
  • Plan B: Source genuine time series data for a hypothetical scenario
  • Plan C: Use share price data from Yahoo Finance

Select a focused business problem:

  • Choose something small, creative, and significant
  • Examples: YouTube channel views, grocery store wastage, employee retention rates
  • Avoid overly complex problems - simplicity with depth is key

Define forecasting relevance:

  • Clearly articulate how forecasting solves the specific business problem
  • Identify tangible benefits (cost reduction, improved efficiency, better planning)

Step 2: Source Your Time Series Data

For Plan A:

  • Request permission to use company data
  • Consider scaling/perturbing data if confidentiality is an issue
  • Ensure you have enough historical points (at least 8 periods if annual data)

For Plan B:

  • Source from government open data portals (avoid Kaggle)
  • Consider tourism data, public health statistics, economic indicators
  • Ensure data is recent and relevant to your chosen scenario

For Plan C:

  • Download share prices from Yahoo Finance
  • Consider additional economic indicators for context
  • Plan to use log-returns for stationarity if using VAR/Granger causality

Data preparation:

  • Clean the dataset (handle missing values, outliers)
  • Organize into appropriate time intervals
  • Document data source meticulously for referencing

PHASE 2: ANALYSIS AND TECHNIQUE SELECTION

▼

Step 3: Perform Initial Data Analysis

Descriptive statistics:

  • Calculate mean, median, standard deviation
  • Identify min/max values and outliers
  • Generate summary statistics

Time series visualization:

  • Plot the full time series
  • Look for obvious patterns (trends, seasonality, cycles)
  • Create decomposition plots if appropriate

Stationarity testing:

  • Run augmented Dickey-Fuller test if necessary
  • Determine if differencing is required
  • Document all findings with clear interpretations

Step 4: Select Appropriate Forecasting Technique

Match technique to data characteristics:

  • Trend but no seasonality: Holt's linear trend
  • Trend and seasonality: Holt-Winters
  • Multiple related time series: VAR (Vector Autoregression)
  • Complex patterns: Prophet with external predictors

Justify your selection:

  • Explain why the chosen technique is appropriate for this specific business problem
  • Connect technique selection to data characteristics
  • Consider business constraints and requirements

Implement the forecasting model:

  • Apply the selected technique to your data
  • Generate forecasts for an appropriate future period
  • Calculate accuracy metrics (MAPE, RMSE, MAE)

PHASE 3: CREATING YOUR PRESENTATION (Part A)

▼

Step 5: Design Your Presentation Slides

Slides 1-2: Company & Business Problem

  • Provide clear company background and industry context
  • State the specific business problem requiring forecasting
  • Make the problem statement compelling and focused

Slides 3-4: Role of Forecasting

  • Explain specifically how forecasting addresses the problem
  • Highlight expected benefits with quantifiable metrics if possible
  • Show the connection between forecasting and improved decision-making

Slides 5-6: Data Overview

  • Present your dataset with key variables clearly labeled
  • Include visualization of the time series
  • Explain data source and relevance to the problem

Design tips:

  • Use clean, professional design with minimal text
  • Include compelling visuals (graphs, charts)
  • Ensure readability with appropriate font sizes
  • Practice your delivery for timing (3-8 minutes)

PHASE 4: WRITING YOUR REPORT (Part B)

▼

Step 6: Draft Your 1000-Word Report

Organization Overview (approximately 200 words):

  • Identify the organization and your connection to it
  • Describe industry context and key challenges
  • Explain decision-making processes relevant to forecasting

Importance of Forecasting (approximately 150 words):

  • Justify forecasting's value for this specific organization
  • Discuss how it enhances decision-making processes
  • Link to efficiency improvements or financial performance

Time Series Data (approximately 200 words):

  • Describe dataset in detail (variables, units, timeframe)
  • Include source information and collection methodology
  • Present descriptive analysis with key findings
  • Include visual representation of the time series

Recommended Forecasting Technique (approximately 200 words):

  • Propose and thoroughly justify your chosen technique
  • Explain why it's appropriate for this specific business problem
  • If using multiple time series, explain the relationship modeling

Results & Analysis (approximately 250 words):

  • Present forecasting results clearly
  • Include visualizations comparing actual vs. forecast values
  • Discuss error metrics and forecast accuracy
  • Provide insightful interpretation of results

Business Benefits (approximately 200 words):

  • Discuss tangible benefits for the organization
  • Consider both financial impact (ROI) and operational improvements
  • Link benefits directly to the original business problem

Step 7: Finalize and Polish

References:

  • Include at least four relevant, recent references
  • Ensure each reference is directly linked to specific points in your report
  • Follow appropriate citation style (likely APA or Harvard)

Format and proofread:

  • Check word count (1000 words ±10%)
  • Ensure all tables and figures have appropriate captions
  • Proofread for grammar, spelling, and clarity
  • Format document professionally with clear headings

Submission preparation:

  • Prepare your Microsoft Word file for Turnitin
  • Zip your data files and forecasting model files
  • Double-check submission requirements and deadlines

AI usage documentation (if used):

  • Document all AI prompts and responses in an appendix
  • Reference AI use appropriately in the main text
  • Ensure original thinking is evident throughout

COMMON PITFALLS TO AVOID

Scope issues:

Data problems:

Technical issues:

Presentation issues:

Report issues: