Dashboard guide

This dashboard analyzes the Climate Change Dataset with attention to data cleaning, missing-value treatment, and visual best practices.

Because climate variables are seasonal, missing numeric values were imputed using the mean within each month (for example, missing July temperature values were filled using the average of other July values). If a month had no available value for a variable, the overall mean for that variable was used as a backup.

The design follows common data visualization principles emphasized in course readings: use position and length for accurate comparison, reduce chartjunk, prefer readable labels, and add interpretation directly next to the visuals.

Summary

References to course best practices

This dashboard applies several visualization ideas commonly emphasized in data visualization courses: Tufte's focus on clarity and high data-ink ratio, Cleveland and McGill's evidence that position and aligned scales support accurate comparison, and Wilke/Cairo-style guidance on restrained color, thoughtful annotation, and explanatory titles.

That is why the dashboard relies mainly on line charts, scatterplots, and ordered bar charts with limited decoration and direct interpretive text.

Total observations

Variables imputed

Values imputed

Temperature trends over time

Carbon dioxide concentration over time

Missing values before imputation

Missing values after imputation

Missing-data handling notes

The dataset contains missing values and a few non-standard placeholders such as "Unknown" and 99999.
To handle these responsibly:

• text placeholders were converted to missing values first
• impossible placeholder values like 99999 were also treated as missing
• missing numeric values were imputed using the mean within month
• if a month-specific mean was unavailable, the overall mean was used as a fallback

Using monthly means is more appropriate than a single grand mean because climate variables are seasonal. For example, replacing a missing winter temperature with the overall annual mean would distort the seasonal pattern.

Missing-value summary table

Seasonal patterns by month

Sea surface temperature vs. air temperature

Correlation heatmap

Interpretation

Several patterns stand out. The temperature lines suggest strong seasonal swings, while the CO2 series shows a general increase across the time span. The monthly summary reinforces that climate variables are seasonal, which supports the choice to impute within month rather than across the entire dataset. The scatterplot suggests a positive relationship between sea-surface temperature and air temperature, and the correlation heatmap helps identify broader associations among temperature, humidity, radiation, and atmospheric indicators.

Recent records after cleaning

Notes for interpretation

This dashboard emphasizes readable comparisons instead of decorative complexity. That choice reflects visualization best practices discussed in many introductory readings: use clear scales, direct titles, and simple encodings that let the data stand out. If you want to adapt this further for your class, you can also add a short final paragraph naming the exact readings assigned in your semester.