Data and Analytic Services

Active

Contract Opportunity

Request for Information (RFI) / Sources Sought for: Data and Analytic Services

1. Request for Information (RFI) is issued for informational purposes and market research only and does not constitute a solicitation. The Government will NOT reimburse any company or individual for any expenses associated with preparing or submitting information in response to this posting.

2. Interested potential offerors are encouraged to respond to this notice. The response shall not exceed 10 pages (including cover page). Describe your organization’s technical capability regarding the tasks identified below. The Government will be reviewing interest and demonstrable experience/ability to propose on a Data Analysis, Artificial Intelligence and Machine Learning capability. RFI submissions will be reviewed on the response to the requirements and tasks contained in the RFI.

************************************************************

Requirements:

The Intelligence Systems Support Office (ISSO), SAF/OC/CDMR, requires data analysis in support of ISSO projects, including the development of algorithms and visualizations from the data. Visualization will be a dashboard documenting behavior detected in analyzed data based on an ongoing initiative to a specific project use case. The dashboards should allow user input to create analytics use cases. Database design and optimization is required to accommodate novel use cases. Part of the process of building the dashboards requires the testing and development of innovative methods for interactive mapping of large quantities of data on the order of 2 million points visualized in an interactive, zoomable, pannable map created by Leaflet and deck.gl, utilizing GeoArrow. The dashboards require development of Extract, Transform, Load (ETL) processes to ingest data based on on-the-fly user-defined requirements.

ISSO requires Machine Learning (ML) analysis to power the dashboards with the following requirements:

1. Computing clusters of points within physical proximity of each other. This comparison must be able to be performed on over 1 million points in under 1 minute.


2. Perform geographic clustering of identified point centroids to determine clusters. These clusters must be converted into geographic polygons.


3. Perform social network clustering of points found to be in close proximity of each other.


4. Model, using an ensemble of time series forecasting algorithms, usual activity in an area of interest (AOI) and signal an alert when actual activity deviates outside an acceptable range. The modeling process must take no longer than 5 minutes to calculate

The data ingestion process for the dashboards requires consuming commercially available data or ingesting from USG-provided Application Programming Interfaces (APIs) on Top Secret (TS) networks. The data must be stored in an open source database with optimizations for geospatial and temporal data.

These algorithms, dashboards and reports will be deployed in containers onto secure networks that support the Defense Intelligence Enterprise. Furthermore, allowing all Department of Defense (DOD) users access, without additional cost to the USG.

ISSO requires the installation of and maintenance of the Posit software suite (Posit Workbench, Posit Connect and Posit Package Manager) to facilitate data science workloads. The software must integrate with Keycloak for user authentication and be installed on Non-classified Internet Protocol Router Network (NIPR) and Joint Worldwide Intelligence Communications System (JWICS).

ISSO requires training for its staff in the R programming language as well as ML and Artificial Intelligence (AI) practices, to encourage automated reporting, rigorous analysis and data-driven decision making. The material should be similar in content and quality to classes at Columbia Business School and Princeton University. The content should be accompanied by corresponding readings from published books on the subject. The training will consist of at least the following topics.

1. Ingest multiple data files into a single data entity using Arrow and the readr package.


2. Ingest data from Excel spreadsheets.


3. Ingest Json and Geojson data.


4. Utilize dplyr and arrow to perform data manipulation such as filtering, creating new columns, summarizing data.


5. Data visualization such as scatterplots, histograms, bar charts, small multiples, line charts and smoothing curves using ggplot2.


6. Generalized Linear Modeling.


7. Fitting boosted trees.


8. Cross-validation.


9. Automating machine learning pipelines with the tidy models suite

a. recipes

b. parsnip

c. workflows

d. yardstick

e. dials

f. tune

g. rsample

10. Automated reporting with quarto.

11. Slideshow generation with quarto.

12. Consuming APIs with httr2.

13. Perform analysis on large datasets using Arrow and DuckDB.

14. Other topics as determined by ISSO.

ISSO requires exposing advanced geospatial capabilities, utilizing modern, open source tooling—such as PostGIS, geos, DuckDB Spatial and simple features in R—to end users with no coding ability. This is required to be in the form of a dashboard that allows users to upload Geojson files to be manipulated through a Graphical User Interface (GUI). Requirements include:

1. The ability to compute buffers of user-defined distances around polygons on the order of 25,000 polygons in under 15 seconds.


2. The ability to determine the intersection between arbitrary points and polygons on the order of 1,000,000 points within 25,000 polygons in under 5 seconds.


3. The ability to export the results for user download as a Comma Separated Values (CSV) or Geojson.

ISSO requires assistance automating data manipulation and model training with various classified data sets related to different ongoing projects. The data manipulation would be computed in R and/or SQL and the model training with various methods such as penalized regression, boosted trees, forecasting and other types as needed.

************************************************************

3. The submitted documentation becomes the property of the United States Government and will not be returned. No feedback or evaluations will be provided to companies regarding their responses to this notice.

Respondents shall clearly mark proprietary data with the appropriate markings. Any proprietary information received in response to this request will be properly protected from any unauthorized disclosure. Any material that is not marked will be considered publicly-releasable.

When submitting a response, please be aware that the Air Force workforce is supplemented by contracted support services personnel who have signed the same Non-Disclosure Agreements (NDAs).

4. Responses shall be sent via email to the Points of Contact NLT the due date/time listed on this notice.

Attachments/Links