Data Analyst
Overview of the Analysis Process
Add Onboarding info here to walk a user through the data analysis process for a site
Practice Listening to Sound Examples
Soundscapes are complex, with many overlapping sounds. When you analyze each audio clip from an ES site, you’ll try to identify specific sounds, even if they blend together. You will be focusing on listening for birds, insects, weather, and people. To prepare, it’s helpful to practice recognizing these sounds before diving into the main audio clips. Below are some practice clips to help you get familiar with each type of sound.
Weather
Rain & Thunder
Wind & Rain
Bird Sounds
Blue Jay
Cardinal
Mockingbird
Insect Sounds
Crickets #1
Crickets #2
People Sounds
Talking
Vehicle Sounds
Light traffic
Tractor
Analyze 2023 Annular Eclipse Data
The ES team analyzed audio from 17 U.S. locations recorded during the 2023 annular solar eclipse. During the 3-4 minutes of eclipse maximum, some sites showed significant sound changes compared to two non-eclipse days. The computer flagged these changes but can’t explain them—that’s where we need your help!
Please review “How to Analyze” before starting.
Click on a Site Below to Start Your Analysis
Find out how we selected the 2023 sites for further analysis
The ES team used computer programs to analyze 2–3 minutes of audio from the eclipse maximum on October 14, 2023—when the Moon covered the greatest amount of the Sun—and compared it to audio from the same time frame on two non-eclipse days, October 12 and 13, 2023. The computer analysis calculated the difference for each site, using the average of the two non-eclipse days as the norm. Any difference on eclipse day exceeding 2 standard deviations from this average is considered significant. In the spreadsheet below, these significant differences are highlighted in yellow, and these are the sites where we need your help in conducting an in-depth audio analysis.
2023 Annular Eclipse Analysis Spreadsheet
2023 Annular Eclipse Analysis Sites Map
2024 Analysis sites are coming soon. While you wait, we hope you will help with the Annular 2023 Data Analysis!
What is a Soundscape?
TED Talk: Bernie Krause: The voice of the natural world
Recommended Video Watch Time: 00:00-6:69
A soundscape is the mix of all sounds you hear in a particular place, creating the overall “sound picture” of that area. It includes natural sounds, animal sounds, and human-made sounds.
There are three main parts of a soundscape:
- Geophony: These are natural sounds made by non-living things. Examples include wind blowing, rain falling, and waves crashing. Geophony is the background sound of nature that happens without any influence from living creatures or humans.
- Biophony: These are sounds made by living organisms. This includes birds singing, frogs croaking, and insects buzzing. Biophony captures the sounds of wildlife in an area, which can change depending on the time of day or season.
- Anthrophony: These are sounds made by humans. Examples include cars driving, people talking, and machines running. Anthrophony often introduces noise into natural environments, which can affect the natural sounds of geophony and biophony.
Studying soundscapes helps scientists understand the health and changes in an environment, how animals behave, and the effects of human activities. Soundscapes provide a complete picture of an area’s acoustic environment and show how living things interact with their surroundings.
Why Analyze Soundscapes?
Help NASA Capture the Sounds of a Total Solar Eclipse!
Time: 00:00 - 2:42
In eclipse science, soundscapes are used for several reasons:
- Watching Animal Behavior: During a solar eclipse, scientists listen to soundscapes to see how animals react. Many animals act like it’s nighttime when it gets dark during an eclipse. They might change their sounds and movements. By listening before, during, and after an eclipse, scientists learn how different animals respond.
- Studying People’s Reactions: Soundscapes can also capture how people react to an eclipse. You can hear people gathering, their excited voices, and the overall noise of the place. This helps scientists understand how eclipses affect people.
- Environmental Changes: When an eclipse happens, the light and temperature change. This can affect the environment, like the wind and insect noises. Soundscapes help record these changes so scientists can study them.
- Cultural Traditions: In some cultures, eclipses are very important. Soundscapes can capture traditional songs, rituals, and other cultural activities during an eclipse. This helps preserve these traditions for future study.
Overall, soundscapes give scientists a way to study the effects of solar eclipses on nature and people, offering more information than just looking at it with your eyes. In our next lessons, we will explore how these soundscapes are analyzed using tools like spectrograms and power spectral density. These tools help us see and understand the different sounds and their strengths in an environment.
The Computer Analysis Process Explained
To efficiently analyze the huge amount of audio data collected during the 2023 annular eclipse and the 2024 total solar eclipse, the Eclipse Soundscapes (ES) team performs computer-based analysis to identify sites with significant acoustic changes. By determining which sites showed statistically meaningful differences in their soundscapes during the eclipse, the team prioritizes specific audio clips for further review.
One key method used in this analysis is calculating Power Spectral Density (PSD), which measures power distribution across different frequency bands in the soundscape. By comparing the Relative Power Spectral Density (RPSD)—the proportional changes in power across frequencies before, during, and after the eclipse—the team can detect subtle acoustic variations potentially linked to environmental or behavioral shifts. These clips, where shifts in environmental or behavioral sounds are most likely, need the help of NASA volunteer scientists for closer examination. Engaging NASA volunteer scientists in this review process allows us to leverage their insights and expertise, enhancing our understanding of the eclipse’s impact on local soundscapes
Establishing Baseline and Variability for Eclipse Soundscapes Analysis
STEP 1. Calculate Relative PSD for Audio Data on Eclipse and Non-Eclipse Days
- Data Collection: Collect PSD data for audio recorded on the eclipse day and two non-eclipse days. The time frame for each recording is 2-3 minutes of annularity for sites on the eclipse path, or 3 minutes centered on the eclipse maximum for off-path sites.
- Relative PSD Calculation: For each site, and each time-period, calculate the relative PSD by normalizing PSD values to a baseline, the average PSD of the two non-eclipse days.
STEP 2. Calculate the Average Relative PSD and Standard Deviation for Non-Eclipse Days
- Baseline PSD: Compute the average relative PSD across the two non-eclipse days at each frequency.
- Standard Deviation: Calculate the standard deviation of the relative PSDs from the two non-eclipse days to measure natural variability in soundscape data.
Example Spreadsheet Title
explanation of this example spreadsheet
Terms and Definitions
Use a Site's Daily Data for Site Comparison Analysis
STEP 3. Identify Maximum Relative PSD Frequency on Eclipse Day
Column I
- Maximum PSD Frequency: Locate the frequency where the eclipse day shows the highest relative PSD difference. (This frequency is listed in Column I.)
Terms and Definitions
STEP 4. Establish the PSD Baseline at Maximum Frequency (Baseline Calculation)
Columns J, K, & L
- Find the PSD values at the maximum frequency from two days prior (Column J) and one day prior (Column K) to the eclipse.
- Calculate the average PSD of these two non-eclipse days and record it in Column L.
- Calculate the standard deviation of the non-eclipse days’ PSD values, placing it in Column M.
STEP 5. Calculate Relative PSDs
Columns N,O,& P
- Relative Standard Error: Calculate the relative standard error by dividing the standard deviation in Column M by the average PSD in Column L (Column N).
- Relative Average PSD: Set the relative average PSD for non-eclipse days to 1 (Column O).
- Relative PSD on Eclipse Day: Record the maximum relative PSD for the eclipse day in Column P.
- Threshold Comparison: To determine the significance, divide the eclipse day relative PSD (Column F) by the relative standard error (Column N) and record in Column Q. If this value exceeds the 2σ threshold, the difference is statistically significant.
STEP 6. Calculate Relative PSDs
Columns N,O,P, & Q
To determine the significance, divide the eclipse day relative PSD (Column F) by the relative standard error (Column N) and record in Column Q. If this value exceeds the 2σ threshold, the difference is statistically significant.
Example Spreadsheet Title
Terms and Definitions
Threshold refers to the level used to determine statistical significance, typically set as 2 times the standard deviation (2σ) above the baseline. This threshold indicates the point beyond which any observed difference on the eclipse day is considered statistically significant.
Document and Visualize Results
- Record the frequency, observed relative PSD difference, and the significance threshold at each relevant frequency.
- Plot the relative PSD differences across frequencies for the eclipse day compared to non-eclipse days, marking the 2σ threshold to visually confirm significance at the frequency of interest.
Example GRAPH Title
EXAMPLE graph here
Text explanation of how to read graph
Interpret Findings - Engage in further analysis
Next Step ES Data Analyst Role
If a significant difference is confirmed, analyze what this deviation at the specific frequency might suggest, such as environmental or behavioral responses to the eclipse.