Ecological Network Testing: Data & Tools

---

PURPOSE

This purpose of the data and tools provided in this repository is to provide insight into the dyanmic coupling between species pairs on the basis of ecological network distance.

It has been designed in a way that it can be modified and run to further analyze the data in ways which I have not, or to gain further insight into my findings.

DATA

There are 3 sets of inital data that are important to understand regarding this project:

ABMI Data

This data is from ABMI, compiled by ABMI's Information Coordinator, Brett Bodeux. It is a set of data from ABMI's Autonomous Recording Units's located in the boreal region of Alberta, and other accessory data.

Interaction Data

This data is from A Trophic and Non-Trophic Seasonal Interaction Network Reveals Potential Management Units and Functionally Important Species. It is a collection of all noted interactions between species across the boreal region of North America.

Boreal Alberta Species Data

This data is compiled by myself. It is a list of all species in the boreal region of Alberta. The species are separated into multiple groups:

• 0: Not present in the boreal region of Alberta in the summer.
• 1: present in the boreal region of Alberta in the summer.
• 2: migratory over boreal region of Alberta, so not necessarily present within sampling months or for extended time.
• 3: grouped species for use with the Interaction data previously mentioned.

WALKTHROUGH

To be successful in this walkthrough, SQLite3 (ver. >= 3.49) and Python (ver. >= 3.12) must be installed on your machine. Also, ensure that all of the files in the zipped init_data folder are extracted and placed in unzipped init_data folder. To prepare the data for analysis, the following commands must be run a terminal whose current working directory is this folder:

Firstly, create a database and upload the inital information:

• sqlite3 data_analysis.db to enter SQLite3's command interface
• .import 'init_data/abmi.csv' Abmi --csv
• .import 'init_data/boreal_alberta.csv' BorealAlberta --csv
• .import 'init_data/interaction_data.csv' Interactions --csv

Next, we need to use the given data to create species pairs found in the boreal region of Alberta and prepare to assign Spearman's correlations to them:

• .read 'sql_queries/species_pair_interactions.sql'
  • Note: remove the last query in this script to calculate a trophic network, as opposed to a combined network
• .read 'sql_queries/species_pair_counts.sql'
• .quit to exit SQLite3's command interface
• sqlite3 -header -csv data_analysis.db 'SELECT * FROM FilteredSpeciesPairs;' > 'generated_data/pair_counts.csv'

Now, we will assign Spearman's correlations to the species counts we have generated and add it to the database:

• python spearmans.py
  • Note: this process takes time. Allow the program multiple minutes to run.
• sqlite3 data_analysis.db to enter SQLite3's command interface
• .import 'generated_data/pair_spearmans.csv' PairSpearmans --csv

Finally, we will assign the uncalculated network distances to this data:

• .read 'sql_queries/species_network.sql'
• .quit to exit SQLite3's command interface
• sqlite3 -header -csv data_analysis.db 'SELECT * FROM SpearmansDistancePairs;' > 'generated_data/interaction_and_spearmans.csv'
• python network.py

Which ouputs a file located in the generated_data folder called completed_network_data.csv which contains all species pairs in the boreal region of Alberta, their Spearman's correlations, and ecological interaction network distance.

FUTURE RESEARCH

CORRELATIONS OVER TIME

An interesting application of this reserach would be to recompute the Spearman's correlations over time. This might help to account for some of the noise seen in my research. I was not able to do this because ABMI did not have the data of repeat visits at particular sites, though I have been told this ABMI is working to collect this data.

In preparation for this data, I have included some queries that would help a future researcher quickly get calculations:

• sql_queries/future_queries/species_pair_counts_over_time.sql would replace sql_queries/species_pair_counts in the walk through to group the data by sites instead of years
• in spearmans.py, swapping the filtering and grouping variable would allow of Spearman's to be calculated over time.
• sql_queries/future_queries/species_network_over_time.sql would replace sql_queries/species_network.sql in the walk through to combine this new data we have just calculated

INDIVIDUAL SPECIES COMPARISON

Another interesting application of this research would be analyzing the relationships between the counts of different species specifically. Doing this with predatory-prey species might reveal some interesting patterns to be analyzed further.

In preparation for this, I have included some queries that would help a future reseacher quickly get data for this analysis:

• sql_queries/future_queries/specific_species_count.sql can be run (see the walkthrough to learn how to run this query) to generate these counts. The researcher must put the species they have chosen to analyze in the IN() function within the query.
• Then, use sqlite3 -header -csv data_analysis.db 'SELECT * FROM SpeciesCounts;' > 'generated_data/specific_counts.csv' to generate generated_data/specific_counts.csv which has the required information.