1. Which dataset did you use?

I used the “Data Center Cold Source Control Dataset” from Kaggle. This dataset represents the operational environment of a modern data center cooling system, focusing on the control of cold source components such as chillers and air handling units (AHUs). It contains 3,498 hourly time-series records, capturing key parameters including:

• Server workload (%)
• Inlet and outlet temperatures (°C)
• Ambient temperature (°C)
• Cooling unit power consumption (kW)
• Chiller and AHU usage levels (%)
• Energy cost
• Temperature deviation (°C)

The dataset is particularly valuable because it reflects the real-world trade-off between energy efficiency and thermal stability, which is critical in data center operations.

2. How did you analyze or prepare the data?

To prepare the dataset for analysis, I performed the following steps:

Data Cleaning & Structuring

• Ensured the Timestamp field was correctly formatted as a continuous datetime variable for time-series analysis
• Verified data consistency across all numerical fields
• Handled aggregation appropriately based on variable type (e.g., averages for temperatures and power, sums for cost)

Feature Engineering (Key Step)

I created several calculated fields to enhance analysis:

• Cooling Efficiency = Server Workload / Cooling Power
• Temperature Difference = Outlet Temperature − Inlet Temperature
• Efficiency Score = Cooling Efficiency adjusted by Temperature Deviation
• Operational Flags, including:
  • Overcooling: High cooling power with low temperature deviation
  • Thermal Risk: High temperature deviation

These derived metrics enabled deeper insights into system performance beyond raw data.

Data Transformation

• Standardized percentage-based fields for proper interpretation
• Created thresholds using data-driven approaches (percentiles and domain reasoning) rather than arbitrary values
• Structured the data to support hourly, daily, and monthly analysis levels

Visualization Preparation

• Selected appropriate aggregation methods:
  • Average for continuous variables (temperature, workload, power)
  • Sum for cumulative variables (energy cost)

3. Who is the intended audience for your visualization?

The primary audience includes:

• Data Center Engineers – to monitor system performance and cooling efficiency
• Facility Managers / Operations Teams – to optimize energy usage and reduce operational costs
• Energy Managers / Sustainability Teams – to improve energy efficiency and reduce waste

The dashboard can support decision-makers by providing clear insights into system behavior and optimization opportunities.

4. What is your visualization about, and what question or problem does it address?

This visualization is designed to analyze and optimize the balance between cooling energy consumption and thermal stability in a data center environment.

Problem Addressed

Data centers consume significant energy for cooling. However, inefficient cooling strategies such as overcooling or poor resource allocation, lead to unnecessary energy costs without improving system stability.

What the Visualization Does

The dashboard provides:

• Time-series analysis to show how cooling responds to server workload over time
• Efficiency insights to identify mismatches between cooling power and actual demand
• Strategy comparison to evaluate which cooling control actions perform best
• Anomaly detection to highlight overcooling and thermal risk scenarios
• Cost analysis to reveal high-cost, low-efficiency operating conditions

5. Did you use any Oracle Analytics AI features when building your visualization (ex. AI Assistant)? If so, please describe how they were used.

Yes, AI Assistant for general understanding of the dataset.