soundDB

Background (who am I?)

GeoCorps intern with Davyd in Denali
Installed, maintained, fixed many soundstations
SPLATted and listened
Coded tools for analyzing Denali dataset

Current Workflow & Tools

Excel at processing data from a single site

Data formats are consistent

File structure is consistent

Current Workflow & Tools

Most common data analysis tool:

Current Workflow & Tools

Analyzing one site: easy

Analyzing many sites: tedious copy+paste

Current Workflow & Tools

"Everyone has their own NVSPL reader function"

New scripts often mean new file-reading code

Mostly ad-hoc solutions

Dealing with accessing many files is still taking time and effort

Current Workflow & Tools

Great at processing data from one site routinely

Good at analyzing data from one site routinely

Still working on analyzing data across many, many sites routinely

soundDB

Python library for accessing any subset of the whole natural sounds dataset

soundDB

Once you write this:

You can easily do this:

Example

import soundDB

sites = ["DENAUPST2015", "DENAFANG2013", "DENAWEBU2009"]

srcids = soundDB.srcid.all(sites)
# now the contents of the SRCID files for those three sites
# are loaded into one DataFrame, called srcids

# analyze as you wish

Example

nvspls = soundDB.nvspl.all(sites)
metrics = soundDB.metrics.all(sites)
listening = soundDB.audibility.all(sites)
dailyPAs = soundDB.dailypa.all(sites)
loudevents = soundDB.loudevents.all(sites)

Example

import soundDB

sites = ["DENAUPST2015", "DENAFANG2013", "DENAWEBU2009"]

srcids = soundDB.srcid.all(sites)

srcids.to_excel("allSrcIDs.xls")
# saves all 3 srcID files
# concatenated into one Excel workbook

soundDB

Access data with minimal programming overhead
Reader functions handle inconsistencies in data
Data returned in highly manipulable structure
Integrates with Python scientific computing tools

Metadata-driven

Use metadata to find sites that match some criteria

Load all their data with soundDB

Eventually: integrate with Metadata Database

Analyze

Questions

Would this be used?

By whom? On what data?

Are they really to learn Python, decently well?

Is it better as a standalone tool, or a programming language library?