Editor’s note: On Wednesday 30 July 2014 analysts, executives and technologists gathered at our Analyst Forum, an event established by a partnership between AnalystOne and the United States Geospatial Intelligence Foundation (USGIF) to help share lessons learned across multiple sectors of the economy, including finance, healthcare, law enforcement, emergency response, scientific research, ecommerce, IT, intelligence, and the military.
We had analysts at the event capturing content that we will summarize here over the next few weeks. And we are surveying 280 people who engaged with us to ensure we are capturing the most important results of the event as well as gaps this community believes we should be collectively tackling to improve the state of analysis. All of this will be shaping our reporting here.
In the post below Katie Kennedy provides a high level overview of the event. Stand by for more details.
Bob Gourley
Publisher, Analyst One
30 July Analyst Forum: The First Report/Summary
Analytics 2014: Insights for Mission impact began with opening remarks from Keith Masback, Chief Executive Officer, United States Geospatial Intelligence Foundation (USGIF). Keith welcomed all and helped bring focus to some of the opportunities the event could help the community achieve. Keith has been an analyst and a leader of analysts and is now a key champion of mission focused capabilities including those supporting the analytical community. He brings knowledge of technology and methodology and critical important missions to this topic. His key message in introducing the event was the criticality of the thoughts that would be exchanged today.
Bob Gourley welcomed attendees. Early on he underscored the gratitude the event organizers had for the sponsors. These firms sponsored the event because of their respect for the mission and understanding of the importance of this effort to the community and they are very much appreciated. Sponsors of the event are: Leidos, BAE Systems, Carahsoft, Cloudera, Digital Reasoning, Digital Globe, IBM and ICG.
Gourley also provided a heartfelt thanks to all attendees for coming, and spent time giving attendees a feel for who else is in the room. Attendees came from the finance sector, media outlets, ecommerce companies, retail companies, the heathcare sector, the scientific research community, law enforcement, plus multiple government agencies including DHS, DoD (Army, Navy, Air Force, Marines ), NRO, DIA, NGA, NIH, NCI, DoT, Treasury, State, VA, HHS, and many others. Commercial firms present included multiple participants from systems integrators with analytical capabilities, high tech analytical tool companies like Recorded Future,WayIn, Pentaho, Optensity and Hexis Cyber.
Speakers at the event are all interesting with great backgrounds in analysis. Speakers included: Tom Lash, Jeff Jonas, Ilkay Altintas, Dave Warner, David Bray, Kirk Borne, Phil Bourne, Bob Grossman, John Harer, Kelly McCue, David Roberts, Abe Usher, Bill Wall, Carmen Medina, Ed Mornston, Mark Ashwell, William Nolte, Matt Devost, Adam Elkus, Ned Moran, Erin Simpson, Dave Gauthier, Bob Jimenez, Jason Thomas, and Scott Sorensen,
Tom Lash of Leidos, a key community thought leader on topics of analytics provided a welcome and an introduction to Jeff Jonas, our first morning discussant. Jeff is widely known in the analytical community for his ability to generate thoughts of direct and impactful relevance to helping people get their missions done. He is also highly regarded as a creator of both concepts and technology that add real value to organizations. Jonas has directed the design and development of multiple innovative systems, including solutions which have countered fraud in the gaming industry (See Jeff’s bio here).
Jonas prepared visuals that paired incredibly well with the messages in his presentation. It was a great help to expanding our thoughts to get Jeff’s perspectives on topics like detection of asteroids and then further turning observations into predictions of asteroid route and event predictions of future observable events like asteroid collisions. This important topic was used to underscore the importance of context when doing analysis. This theme of context was critically important to Jeff’s presentation and was also a recurring theme hit upon many other times throughout the day.
Jeff’s approach is to seek not just a little data, but as much data as possible. And of course to seek and leverage context. There needs to be more data to make better predictions, as well as filling in the gaps where there is not enough data. Jonas used the example of twins. If you have two people who look identical in every way and every feature determines them as one entity, how can you distinguish them? Therein lies the fact that two identical things cannot occupy the same space at the same time. To distinguish the two, there has to be data that supports the case that they are indeed two separate people. The more data gathered, the easier it would be to distinguish the twins and find they are two separate and distinct individuals.
By the way, Jeff received thunderous applause. It was great having him interacting with us all throughout the day to help underscore the lessons he left us with.
Bob introduced the next speaker, Dr. Ilkay Altintas, the Director for the Center of Excellence in Workflows for Data Science at the San Diego Supercomputer Center (SDSC), UCSD. Altintas addressed the question of “why data science workflows?” She answered, “workflows contain a lot of small steps that become programmable and reproducible scalability.” A project the SDSC has been successfully working on is entitled “WIFIRE”, a scalable data-driven monitoring, dynamic prediction and resilience cyberinfrastructure for predicting and monitoring wildfires. Wildfires are difficult to predict, therefore WIFIRE supports an integrated process that analyzes wildfires, incorporating observations with real-time data.
Bob asked Dr. Altintas to discuss “The Scientific Method.” Here is the idea we pondered: almost all of us grew up learning “The Scientific Method”. The scientific method of observation, hypothesis, prediction, experimenting is still critically important. But because of new technologies available to researchers and the incredible amount of data available, it is no longer the only workflow available to construct workable models of the world or to advance science and understanding. Dr. Altintas provided context on this topic that underscored the importance of considering workflows in analytical processes at any sector.
Bob then introduced Dr, Dave Warner, M.D., Ph.D., Medical Neuroscientist and the Director of Medical Intelligence at MindTel. Dr. Warner opened with “Computers are rocks that do math and they should worship us”. His idea that “dots are stupid” led to his highly innovative creation of hyper dimensional dots, where an observer can see thousands of different information on the dot, while taking advantage of abstract information. The dots are geolocated allowing the viewer of the dot to hover over an area and instantly have access to densely populated information on the area observed. Creating the models of information preserves relationships and transformation. Each 3-dimensional dot instantly presented multiple facets of data collected in a way that was easily understood by the vast majority of people.
After a short morning break for direct networking, everyone was ushered back into the hall where the next speaker was poised and ready to address the audience. Dr. David A. Bray, Chief Information Officer, FCC, prompted the question of “how to tackle the changing world?” Points he drove home through example after example were that analysts needed enhanced context to ensure optimal assessments. Among the many considerations he helped us think through were the fact that we all need to recognize that you can’t just build higher walls for Internet security. Computers need to alert humans when a threat arises. The FCC is making an application that will send pings during a crisis, constantly updating individuals while simultaneously sending feedback to responders. Innovation is needed more than ever; can public service change quickly enough? It is a technology as well as a people issue. The future includes algorithms working alongside public service workers: when will having software make an unbiased recommendation better than a human? Humans are biased, so maybe algorithms are better for decision-making in some cases. In all cases, context is important.
The next segment included Cross-Sector Analytical Lessons Learned. The moderator was Dr. Kirk Borne, Data Scientist and Professor, George Mason University. Other segment participants included Dr. Philip E. Bourne, Associate Director for Data Science, National Institutes of Health; Dr. Robert Grossman, Director, Center for Data Intensive Science, University of Chicago; and Partner, Open Data Group; and Dr. John Harer, Professor of Mathematics and Computer Science, Duke University. Each individual had their own unique lessons they had learned. First was Dr. Borne, he addressed his lesson with the idea that no matter what field you are in, you can talk with people in other fields when you are speaking the trans-disciplinary language of data science. Second was Dr. Bourne, reasoning that answers to big data problems can come from anywhere either cofounded in a journal, a pandemic modeling article, or a 15 year old who was published in a leading journal. Third was Dr. Grossman, addressing that all men can see tactics, but who can see strategy? Tools are understood, but strategy is not as thought-through. Lastly, Dr. Harer prompted that successful collaborations have understandings of shared work, learning others’ languages and fields, and working with someone who needs you and who wants you to need them. Each panelist emphasized that there were many lessons learned and many more to come.
The next panel was law enforcement, with moderator: Dr. Colleen “Kelly” McCue, Senior Director, Social Science & Quantitative Methods, DigitalGlobe Analytics. The panelists were David J. Roberts, Senior Program Manager, IACP Technology Center, International Association of Chiefs of Police; Abe Usher, Chief Technology Officer, HumanGeo; and Bill Wall, Vice President, Praescient Analytics. Each individual on this panel presented around a central theme that positively correlated better law enforcement and increased data usage. Usher presented on the lessons learned from the London Olympics. He said that as law enforcers, there should be proper identification of observation space and the creation of simple systems for recognizing normal and abnormal behavior. If there is a baseline, then it would be simple to find the abnormal. Next Wall addressed the London riots of 2011, relative to geospatial technology. Wall addressed that there should be geospatial data to build a database of individuals and incidents. Geospatial aspects of big data give us special capabilities to contextualize events. For the London riots, using the geospatial technology, law enforcers were able to find the rioters prior to their insurrection. Next, McCue presented on the Northern Virginia shooting incident that was solved using geospatial predictive analytics providing the high probability target areas that led to the suspect’s arrest. Lastly, Roberts brought to the audiences’ attention the immense impact the poor economy has had on the police workforce. Optimizing resources is extremely important and information-sharing capabilities are essential. Research has emerged looking at crime analysis data to identify systemic problems. Having data can create a mosaic, where many pixels create an image that can be used to solve crimes. The increased gathering of data has led to significant breakthroughs in solving crime.
After a lunch reception and exhibits, the next panel eagerly anticipated the audience’s return to their seats. The next panel was the Education of the Analyst, with Carmen Medina, Specialist Leader, Deloitte Consulting; Ed Mornston, Director, Human Development Directorate, National Geospatial-Intelligence Agency; and Dr. William Nolte, Program Director, Intelligence Center of Academic Excellence, University of Maryland School of Public Policy, and Mr. Mark Ashwell, internationally known developer of analytical strategies who has served as director of intelligence at the Royal Air Force and other UK MoD positions. Medina addressed cognitive diversity and the future of intelligence work, meaning that not all thinkers are the same and there are inherent cognitive differences. She shared the formula calculating what you know as reality – observation error – bias = what you know. Mornston followed with the idea of creating a culture of learning where there is a thirst for innovation, while addressing challenges from declining resources. He urged that agility is key and needs must be anticipated. Nolte pointed out that there are now technically illiterate students and technologists with little worldliness. Analysts must think analytically, not tactically. The 21st century intelligence will be about information, not secrets, like the 20th century. Ashwell stated that the main issue is the challenge of using open source and classified information; analysts must be able to use both to not fall behind. Analysts are not prepared for using big data, we must move from the word to visualization.
In the question and answer session, Dr. Ilkay Altinas asked about the new age of data around individuals and the need for individuals to have data analysis skills and requested panelist thoughts on that topic. This powerful question generated great discussion and is a topic we will continue to help think and write about.
The following panel addressed the analytics of cyber conflict with moderator Matt Devost, President & CEO, FusionX and panelists Adam Elkus, Senior Analyst at Analyst One, Strategic Planner and Ph.D. Student, Computational Social Science, George Mason University; and Ned Moran, Professor, Analyst, Cyber Practitioner. Elkus started the discussion with the notion that cyber conflict takes place in a human made environment and within natural laws. It is a tangled mess of man and machines. What you intend a machine to do is not what it actually does. A lot of cyber issues come down to adaptation and speed. Moran stated that cyber threat intelligence means having better informed defensive decisions. There is a need to get ahead of the attacker through analysis. Moran addressed how we defend against attackers. The landscape is changing, but still the defenders are the ones always responding to the attackers. Moran said that “humans always make mistakes and attackers take advantage”. Cyber conflict is a defenders game and will be for a while.
The final panel was Lessons From and for the National Security Community, with moderator: Dr. Erin M. Simpson, Chief Executive Officer, Caerus Associates and panelists Dave Gauthier, Activity Based Intelligence Portfolio Lead, Analysis Directorate, National Geospatial-Intelligence Agency; Bob Jimenez, Chief Technology Officer, National Reconnaissance Office; and Jason Thomas, Manager of Innovation for the Government, Thomson Reuters. This panel addressed the challenges with data in national security. A few challenges include how data correlates, how to engage an NRO user, and how to recruit people to solve problems. Thomas said that “we have to adapt to what we think is coming, which is not a lateral movement”. In some cases we store more data than necessary and in others, we can’t get our hands on data we need.
The final speaker was Scott Sorensen, Chief Technology Officer, Ancestry.com. His presentation started with a fantastic way to grab the attention of everyone in the room. We all know Ancestry.com and the overview and insights of the technological approaches there (and impact on solutions for customers) really riveted everyone’s attention. “At Ancestry.com, we tell stories from data; we fill in the gaps of family history. With the use of AncestryDNA, we have collected sampled from 500,000 people and found over 10,000,000 fourth cousin matches. We provide context to family history by working with government agencies and churches.” Handwriting recognition locates words on a handwritten document, extracts the words and makes the documents easier to place in family trees. The exciting stories lead directly to very relevant conclusions for any organization seeking to enhance analysis, in any sector. We will provide details on these lessons in coming posts, Scott has agreed to share his graphics with us and they flow directly to those lessons, so stand by for more. The short version, however, is this: How you organize you analytical efforts, including the people working them and even where they sit, is of critical importance. Getting groups to work together is a challenge especially with scientists and software engineers.
The event ended on a high note, and everyone, excited from the day’s events, filed into the adjacent room to mingle and discuss the information learned. The end of day networking allowed for attendees to more deeply connect and interact with each other and speakers.
We would like to conclude this overview with a huge thank you to the team at USGIF who worked tirelessly to ensure this event would come off without a hitch. Their attention to detail and focus on getting the right things done went above and beyond the call of duty