Chronicle your life with data analysis. Tell your story one tiny message at a time.
Storyteller makes it easy to quickly jot down feelings, facts, and events! Use those notes to remember your thoughts about the past, look for patterns in observations, and create interactive visualizations that help tell your stories.
Try out Storyteller and check out some of its features below.
Storyteller is designed to make it fast and easy to jot down notes to your future self. But don't write essays! Those take time. Storyteller's Stories are intentionally limited in length to force you into capturing thoughts while they are fresh.
Type in whatever you want in the Story box. It is automatically time-stamped, and you can specify a duration to associate with the Story. Then hit Keep It! and your Story will be stored. It's like a status update, but only you can see it.
Later on, you can search through your Stories to find specific key words or create visualizations based on your Stories.
In the Story box, all items that are separated by new lines (by hitting the enter key while typing) appear as multiple entries in the same Story.
Just below the Story box is a Search field. Type in any word and Storyteller will highlight Stories containing that word.
Tagging Stories is a useful practice. For example, some Stories are lists of things that need to be done, so writing Todo as the first word of the Story helps to keep them easily reachable.
After Keeping a Story, any items that are on their own lines can be Toggled on and off by clicking on them. Finish an item on your Todo list? Toggle it off! Pick up some items from your grocery list? Toggle them off!
By combining Search and Toggling, you can quickly see what still needs to be done while feeling pleased with all of the things you've already accomplished.
Remember that Duration field in the Story box? Use that to keep track of how long your activities last.
Visualize your Stories to see individual entries as well as how much time you spent in total on a specific activity.
The screenshot here shows hours spent working on the UNO Engineering Forum. The tag 'TS UNOEF' keeps this activity separate from others and makes Visualization a breeze. Notice that the black dots correspond to individual Stories' durations and the gray dots correspond to cumulative durations. Selecting a dot on the graph highlights the Story itself, and scrolling through Stories brings focus to the corresponding dots on the graph.
For work/billing purposes, having a small log of each hour of work can be very beneficial. Also keep track of workout hours, time spent dancing, meals, naps, or anything else you can think of!
Want to track something other than time? No problem! Just write down a number and a unit on the same line, and Storyteller will be able to visualize it for you.
The screenshot here shows a series of entries corresponding to fill-ups at the gas station. Plotting miles is as simple as creating a Cumulative Visualization with the tag "Gas (Miles)" -- Storyteller will find all posts that contain the word "Gas" and then all the numbers that are next to the word "Miles". Other possible visualizations are "Gas ($)" and "Gas (Gallons)". Visualizing "Gas" by itself with no units attached will show the Duration by default.
Each Story is timestamped, and Visualizing the time-of-day of those Stories can be very interesting!
Time of day along the vertical axis can identify daily patterns with ease. Date along the horizontal axis serves as a reminder of blocks of time.
The screenshot shows an overview of work hours tagged by 'TS Unocity', 'TS UNO', 'TS UNOEF', and others '^TS'. It is clear that UNO work was intense and centered in March of 2019, while UNOEF work was equally intense and centered around September of 2019.
In addition to time-of-day, Storyteller can visualize day-of-week plots!
The screenshot shows meeting times with our interns. They're students, so their schedules change every semester. Correspondingly, our meeting times developed a new norm in August. Take note of the shift from Mon/Wed to Tue/Thu.
Storyteller wants to help you tell your life's story. Between fast Story entry and simple-yet-powerful Visualizations, Storyteller provides many tools to you.
But if you need more flexibility than what is offered by Storyteller, then take your data into your own hands and perform any kind of analysis you can imagine. Storyteller can send all of your Stories, their Edits, and their Toggles to you in a simple .xlsx file any time you desire.
Storyteller is a web app, so it runs on any device that has a browser and an internet connection. See your same Stories and Visualizations on a desktop computer, laptop, or phone.
Get extremely fast access to Storyteller on your phone by adding a shortcut as an icon on your home screen. Simply click on "Add to Home Screen" on Android or iOS devices. For Android, this button is in either the tiny house icon on the URL bar or on the three dots that show the page's menu. On iOS, click on the Bookmarks or Sharing icon.
Data Scryer is a high-dimensional visualization tool.
Data Scryer uses state-of-the-art machine learning techniques to perform dimension reduction so that you can quickly see patterns in your data. You can see relationships between different sensors at any moment in time, or you can explore changes in the entire system by analyzing all sensors simultaneously.
Data Scryer's algorithms maintain relationships in the data while reducing them to two dimensions. In two-dimensional space, patterns and clusters are immediately recognizable even though those same clusters in high-dimensional space can be extremely difficult to identify.
Data of high dimensions is trivially easy to generate. Consider a few examples: closing prices for several stocks over a 251-day period results in a 251-dimensional data set; 50,000 metered voltage readings results in a 50,000-dimensional data set; monitoring age, weight, blood pressure, oxygen level, and pulse results in a 5-dimensional data set.
High-dimensional data can be augmented with desired predictions in order to find patterns linking the two together.
Data Scryer can visualize partially-labeled data in order to help you identify clusters and patterns. Partial labeling can take the form of insufficient information (leading to semi-supervised learning) or of a desire to exclude irrelevant information (filtering the data).
Filtering can be seen in the screenshot here. The data set consists of 14-days' worth of stock closing prices for 10,000 stocks augmented with the size/side of their returns 3 days later. Using only past data and selectively highlighting Big Losses and Tiny Losses shows that there are patterns that can separate the two loss sizes even three days in advance.
(Unfortunately, it's incredibly difficult to separate gains from losses three days in advance!)
Different data sets have unique clusters and patterns, and so there is not always a 'best' algorithm to use in order to discover data structure. Data Scryer implements several algorithms and makes it very easy to rapidly explore different algorithm options.
Data Scryer is purely unsupervised, so its calculations ignore predicted labels even if they are provided. Any patterns that emerge are properties of the data itself and are not coerced while learning. This results in fast iteration and helps with exploration even when predictions are not available.
Power system sensors record voltages and currents. It is easy to plot each reading individually as a function of time, and it is just as easy to overlay all of them onto a single plot. However, even with all plots overlaid, it is not possible to see relationships between different sensors. If your system has 50,000 three-phase buses, how are you going to separate and organize all of those sensors' readings? Use Data Scryer.
Unocity provides consulting in the fields of machine learning, data analytics, predictive analytics, power systems, and general electrical engineering.
We rely heavily on automation to solve problems. Often organizations will have data neatly bundled together, but the bundles are scattered across different areas based on the data's usage in the organization. We customize software that bridges gaps, and because we have expertise in engineering, we can create new applications that open novel possibilities in your organization.
Additionally, we provide training. We have a long history of consulting, performing research, and teaching, as can be seen on our capabilities page. We can create a curriculum designed for any skill level between high school and doctoral.
Here we will elaborate on some services that we provide. If you are interested in any of these, or if you have ideas and needs for new ones, feel free to contact us !
If you have a large amount of data and need help making sense of it, then our machine learning knowledge may be useful for you. We can use deep learning for classical problems like regression and classification, but we enjoy cutting-edge problems such as semi-supervised learning, a set of models and constraints that are especially suited to big data. We can consult with you on your specific problem, data, and pipeline in order to obtain the best results.
Additionally, we believe visualization using machine learning is very valuable. Being able to perform extreme and nonlinear dimension reduction on data allows us, as humans, to clearly see relationships between data points that were previously invisible. Trained machine learning models can process new data very quickly, therefore these models can be used for interactive visualization. Real time, interactive visualizations can provide novel yet powerful ways of understanding data and making decisions.
While power system automation can improve safety and protect lives, not all automation needs to involve high-risk situations. Nevertheless, automation can bring enough of a speed increase to make some tasks real time.
The front desk system of any event needs to be fast enough to register new attendees, print name tags for them, and print customized attendance certificates for all. In order to scale, such a system needs to be able to handle multiple printers and paper types flawlessly, and it needs to be able to provide both manual and automatic access to the registrant database and art assets. A small event could handle some of these requirements by hand, but an automated system is absolutely necessary for any larger event.
With our system, automation extends to the artistic vision. A fixed template can be designed for a consistent look, and custom elements can be randomized or procedurally decided. For example, sponsor logos can be arranged randomly for fairness and uniquely for each individual, but company affiliation arrows still are drawn precisely each time. Any additional custom information, such as unique login names or passwords, can be added, too.
Starting from a master database file (including simple ones such as spreadsheet files), our automation can create a fully interactive web version of the schedule as well as a professionally-typeset printed version. Each time the master is changed, the web version can receive the updates immediately without needing to be redesigned.
The print version, too, will receive the updates, but printed materials obviously will need to be reprinted in order to see the updates...
This automation allows all of the information to remain correct and timely. Last-minute changes to the schedule can be included without worry. Designing a printed schedule takes a long time, but that design can be done in advance so that the actual printing can take place as close as possible to the day of the event for the most accurate possible printed schedule.
Unocity provides several services in the power systems area. If you already have a working system but need additional, knowledgeable helpers, we can provide them. If you want a working system but are unsure how to create one, our advice can point you in the right direction. If you had a working system but something went wrong, our consultation can aid the situation. If you want to expand your system to new horizons, our research expertise can show you the way.
As educators and academics, we cherish the students who go through UNO's Power and Energy Research Lab. Over the course of their Master's and PhD degrees, they learn many valuable skills and are exposed to real engineering problems. Unocity as a company employs those skills to help you with any engineering problem you may have.
And just how we trained those students, we can train you or your organization as well.
Sometimes you have a piece of equipment that does not quite work the way you want, and your only clue to fixing it is a Programmable Logic Controller's (PLC) ladder diagram provided to you by the manufacturer. We can help you track down the problem as well as provide advice on how to fix it. Or if you would like to be able to fix it yourself, we can provide training and instruction to improve your skills.
Maybe you want to better integrate different pieces of equipment into your workflow. Hardware PLCs can communicate over the network using the Modbus protocol, but the protocol is not just limited to hardware PLCs. We can create visualizations and dashboards to help you monitor your devices, or we can teach you how to create them yourself. It might be possible to create an overarching flow to your process that no single PLC could control on its own. Our consulting services can help you automate the automation.
Unocity specializes in Learning, Analytics, and Automation. We are a team of researchers, educators, and industry specialists, and all of us are proud of our homes in New Orleans.
Rastin Rastgoufard, Ph.D. created and operates the company Unocity to provide Machine (and Human!) Learning, Data Analytics, and Automation Consulting based on his doctoral research. He completed his Ph.D. in 2018 from the University of New Orleans after receiving his Master's degree from the same university in 2012. He was in Tulane University's last Electrical Engineering graduating class in 2008 where he received his Bachelor of Science.
Rastin has many research publications in machine learning as well as power systems. He has presented his works multiple times internationally. Some of his publications are captured in his Ph.D. Dissertation and Master's Thesis, both of which are freely-available online.
In addition to researching, running Unocity, and creating Storyteller and Data Scryer , Rastin has been on the Organizing Committee of the UNO Engineering Forum since 2014 where he writes software to automate many of the technical issues pertaining to the operation of the Forum, including website management, registration handling, and audio/visual configuration.
Itti earned his B.Eng. in Control Engineering from King Mongkut’s Institute of Technology, Thailand in 1996; Master of Manufacturing Management from the Pennsylvania State University in 2002; Master and Ph.D. in Electrical Engineering and Computer Science from Tulane University in 2007 under Dr. Parviz Rastgoufard’s advising. He worked in petrochemical and software industries. His research interests are power systems optimization and economics, energy efficiency, software development, and business intelligence. He has worked with more than 20 Masters and Ph.D. students. He has been PI, co-PI, and project participant in more than 15 research grants and projects including the Department of Energy Smart Grid Initiative grant and Louisiana Board of Regents' WISE grant.
Before joining Unocity Mark spent 37 ½ years with Entergy, a large electric utility in the middle south region of the USA. During that time Mark spent time as an Engineer, Supervisor, Superintendent, and Manager, in Distribution Design, Construction, Supply Chain, Distribution Substation Engineering, Transmission System Engineering, Transmission Design, and Grid Modernization. In his most recent roles Mark managed development of standards for materials, design and construction of Transmission Systems and Substations; labs for testing protection systems, including hardware, software, firmware, and usage; selection of suppliers and purchase of materials; and, the integration of Transmission information systems with those being developed for Grid Modernization.
Mark studied Electrical Engineering at the University of Michigan and attained his MBA from the University of New Orleans. He is a Professional Engineer in the State of Louisiana.
Parviz Rastgoufard joined the University of New Orleans (UNO) and serves as the Entergy Endowed Chair for Power Systems Engineering – a $2 million endowment established by Louisiana Board of Regents and Entergy Services Inc - since August 2007. He is currently engaged in research and teaching Power Systems Engineering and serves as the Director of Power & Energy Research Laboratory (PERL) and the General Chair of the UNO Engineering Forum since 2009 and chairing panels of experts on Engineering Education, Cybersecurity, and Smart Cities in the years 2016, 2017, and 2019 respectively.
Before joining UNO, he served in many roles at Tulane University from 1987 to 2007, including Professor and Chair of Electrical Engineering and Computer Science (1998 – 2007); Director of Electric Power Research Laboratory (1991 – 2007); Chair of American Association of University Professors (2003 – 2007); member of Tulane Senate for six years, and Entergy Endowed Chair in Electric Power Engineering (1998 –2007). Through collaboration between Entergy and Tulane School of Business and School of Engineering, he was instrumental in creation of Entergy-Tulane Energy Institute in 2004-2007. He served as General Chair of Society for Electric Power Research Implementation (SEPRI), a collaboration between universities and industry in State of Louisiana, during 1990-1994, and the co-founder of Tulane Engineering Forum and serving as the member of its organizing committee during 1998-2004. Parviz serves as the Emeritus Professor of Electrical Engineering and Computer Science and Emeritus Entergy Endowed Chair for Electric Power Engineering at Tulane University since July 2007.
Prior to joining Tulane University, he was a member of faculty of Department of Electrical and Computer Engineering (ECE) at North Carolina State University in the years 1983-1987 and founding member of Electric Power Research Center - established by Department of Nuclear Engineering and ECE with membership from electric utilities, government agencies, and Electric Power Research Institute (EPRI).
Parviz has served as Principal Investigator on more than 80 funded grants and projects resolving issues related to Power Systems Engineering reliability, planning, operation, safety, prevention of blackouts, hardening of transmission network, system resiliency, smart grid, cybersecurity, and hurricane preparedness of industrial plants in Gulf States. Through the funded grants and projects, Parviz successfully created the Master and Ph.D. programs in Power Systems Engineering at Tulane University and University of New Orleans resulting in more than 75 Master and Ph.D. students including 4 post-doctoral fellows. Several of his students have leading roles in different sectors of workforce nationally and internationally including universities, government, and industry. Parviz has served as Principal Consultant on more than 20 projects.
Parviz is a senior member of IEEE and a member of New Orleans ACE Board, and has served as professional engineer, member of industry Board of Directors, member or chair of 25 Search Committees, and as research and engineering consultant for several local and national companies since 1988. Dr. Rastgoufard received his Bachelor of Science in Electrical Engineering from State University of New York in Buffalo in 1976, and M.S. and Ph.D. in Systems Science from Michigan State University in 1978 and 1983, respectively.
Parviz Rastgoufard and our colleagues worked on 100 projects over the last 30 years, and we now bring our skills to Unocity. We have trained more than 75 graduate students (Master's and Ph.D.) and many more undergraduate students serving the professional engineering community.
The following is a list of some of our recent power systems project areas.
Real Time Modeling and Simulation of Electric Power Systems