IN-SPIRE Training Office of Portfolio Analysis Division of Program - PowerPoint PPT Presentation

IN-SPIRE Training Office of Portfolio Analysis Division of Program Coordination, Planning, and Strategic Initiatives National Institutes of Health Todays Instructors: Patricia Forcinito and Matt Perkins OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

Mission of the Office of Portfolio Analysis OPA Director: Dr. George Santangelo Facilitate Data Develop Cleaning & Analysis New Methods Support Build Consult & data-driven Tools Collaborate decision-making Disseminate Best Practices Classroom Training (Custom Classes Available) Online Training Web Resources (Case Studies, FAQs) Office Hours Symposia https://dpcpsi.nih.gov/opa/index OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

iSearch iSearch is NIH’s next -generation portfolio analysis platform, providing comprehensive, easy-to-use access to a carefully curated, extensively linked data set of global grants, patents, publications, clinical trials, and drugs. https://itools.od.nih.gov/dashboard/#login OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

Course Objectives Part 1 • Understand how IN-SPIRE works and what it might be used for. • Learn how to import a dataset, do some visualizations and export results. Part 2 • Learn how to use the following features: • Searches • Search Networks • Time Tool • Correlation • Facets OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

About IN-SPIRE • Developed by Pacific Northwest National Laboratory (PNNL) to help analyze huge datasets • Can help understand data in two different ways: – Convergent • Focusing on a particular set of information or relationships. – Divergent • Remaining open to potentially interesting information or relationships. • Add efficiency to the process of analyzing large datasets. OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

Examples of IN-SPIRE use in portfolio analysis OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

Is there overlap between agencies/ICs/divisions? IN-SPIRE Galaxy View 1 2 3 5 6 7 4 Applicable for other comparisons, 9 8 eg: • between NIH & non-NIH grants 14 10 11 • Different NIH POs working in the 16 same research area etc. 13 12 15 20 17 18 IC (b) 19 IC (a)

Evolution of Portfolios: Stem Cell Research Year 2009 IN-SPIRE ThemeView Classic brain, nscs, nsc Human, gene, embryonic 291 Projects Searched QVR f or “S tem Cell ” in Title and Abstract OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

Evolution of Portfolios: Stem Cell Research Year 2013 IN-SPIRE ThemeView Classic Neurons, brain, neural stem cells 193 Projects OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

Identify training activities in T34 grants • Identify Portfolio. • Use Search Networks to combine range of Search Terms and group to create categories. • Develop understanding of training activities. OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

Understanding datasets / Manual Coding OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

How does IN-SPIRE create a visualization with my documents? Source Dataset i.e. grants, pubs, etc. Text to be mined by IN-SPIRE Specific Title Abstract Aims Today’s example: Excel Spreadsheet with grants belonging to the transplantation portfolio OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

How does IN-SPIRE create a visualization with my documents? Source Dataset i.e. grants, pubs, etc. Signature clustering projection scan generation Clusters are Uses these IN-SPIRE groups Identifies the terms arranged in a distinguishing the documents into that are least high- words to create a clusters based on randomly dimensional signature for each their signature space and then distributed document similarity . flattened down to a 2D graph Determines distinguishing words called “Major Terms” OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

IN-SPIRE Visualization Galaxy View • Each document is represented by a dot or “star” in the “sky” • Documents are cluster together according to their content similarity OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

IN-SPIRE Visualization ThemeView Classic 3-D representation of the clusters • By default the peak height represents number of documents • Peak height can be adjusted to represent for example: total awarded $, etc. OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

Demonstration followed by hands on exercise OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

Part 2: Objectives • Learn: • Searching • Groups • Search networks • Time tool: documents, groups, flows • Correlation tool • Facets tool OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

Search Networks • Similar to searches / groups. • Two advantages: • Useful for combining a number of searches and visualizing results. • Search networks can be saved and copied to other projects or shared between users • Can use searches created in search function (and groups) or create new ones. • Links between searches are easy to create and remove. • Values update when new data is loaded. • OPA guidance here: https://opa- trainingportal.od.nih.gov/?page_id=961 OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

Sharing Search Networks • Once you’ve created a Search Network, it will appear in the project folder, and can be dragged to other projects to use on other datasets. • To transfer Search Networks to another user: • Copy to Home folder • Log out of IN-SPIRE. • Use the IN-SPIRE Meta Data Sharing tool (likely to be at): • C:\Users\perkinsmj\INSPIRE\bin\Meta Data Sharing • Follow the process in reverse to import. OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

How do I get access to the tool? http://in-spire.pnnl.gov/government.stm OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

Leave empty OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis

During today’s class we covered: • How IN-SPIRE works and what it might be used for • How to import a dataset (differences within computation, categorical, date, and numeric fields) • How to get IN-SPIRE main visualizations (Galaxy View and Theme View Classic) • Used: • Search feature • Created groups using multiple methods • Search networks • Time tool • Correlation tool • Facets tool QUESTIONS? OPA_T#1192_Dec-13-2016 Office of Portfolio Analysis