open share links close share links

Like financial assets, data assets can have different levels of liquidity. Certificates of deposit tie up money for a certain length of time, preventing other use of the funds. Siloed business applications tie up data, which makes it difficult, even impossible, to use that data in other ways across an organization.

A recent research briefing, Build Data Liquidity to Accelerate Data Monetization, defines data liquidity as ease of data asset reuse and recombination. The briefing was written by Barbara Wixom, principal research scientist at the MIT Center for Information Systems Research (CISR) and Gabriele Piccoli of Louisiana State University and the University of Pavia.

Unlike physical capital assets listed on a corporate balance sheet, such as buildings and equipment, data does not deteriorate over time. In fact, it can become more valuable as it is used in different ways.

While data is inherently reusable and recombinable, an organization must activate these characteristics by creating strategic data assets and building out its data monetization capabilities, the authors explain.

Typically, companies use data in linear value-creation cycles, where data is trapped in silos and local business processes. Over time, the data becomes incomplete, inaccurate, and poorly classified or defined.

To increase data liquidity, organizations need to decontextualize the data, divorcing it from a specific condition or context. The authors suggest using best practices in data management including metadata management, data integration and taxonomy/ontology to ensure each data asset is accurate, complete, current, standardized, searchable, and understandable throughout the enterprise.

Such data management practices build key enterprise capabilities like data platform, data science, acceptable data use, and customer understanding, which increases datas monetization potential.

As a companys strategic data assets become more highly liquid and their number grows, data is made increasingly available for conversion to value, and the companys data monetization accelerates, write the authors.

In explaining how an organization can create highly liquid data assets for use across an enterprise, the authors cite the example of Fidelity Investments, a Boston-based financial services company.

The firm is combining more than 100 data warehouses and analytics stores into one common analytics platform, built upon five foundational structures:

Fidelitys goal is to organize the data around a common set of priorities such as customer, employee, and investible security. The result will be strategic data assets that are integrated and easily consumable. We want to create long-term data assets for creating value, not only immediately, but also for use cases that are yet to be identified, says Mihir Shah, Fidelitys enterprise head of data, in the briefing.

As long as Fidelitys internal data consumers follow core rules, they can combine data from different sources and build for specific requirements. Not only has Fidelity already created valuable data assets through this platform, it has begun to identify value-add opportunities using data that were never before possible activities that would add value for customers, revenue, and efficiency, according to the authors.

Once data is highly liquid, future ready companies can use it to produce value in three ways, according to MIT CISR research:

Fidelity is one of more than 70 strategic data asset initiatives that MIT CISR researchers uncovered in the course of interviews with its member organizations. The projects illustrate how the beauty lies not in a single use of data, but in the recurring reuse and recombination of the carefully curated underlying strategic data assets, the authors write.

As companies transform into future-ready entities, they need to view their strategic digital initiatives not simply as a way to exploit digital possibilities, but also as opportunities for reshaping their data into highly liquid strategic data assets, they conclude.

Read the original here:

Liquidity is key to unlocking the value in data, researchers say - MIT Sloan News

In recent years, immune-based treatments for cancer have buoyed the hopes of doctors and patients alike. Drugs called immune checkpoint inhibitors have provided lifesaving benefits to a growing list of people with several types of cancer, including melanoma, lung cancer, bladder cancer, and many more.

Despite the excitement surrounding these medications, a frustrating sticking point has been the inability of doctors to predict who will benefit from them and who will not.

On August 25, 2021, a group of researchers from Memorial Sloan Kettering Cancer Center reported in the journal Science Translational Medicine that a specific pattern, or signature, of markers on immune cells in the blood is a likely biomarker of response to checkpoint immunotherapy. Within this immune signature, a molecule LAG-3 provided key information identifying patients with poorer outcomes.

This link was discovered in a group of patients with metastatic melanoma and validated in a second group of patients with metastatic bladder cancer, suggesting that this potential biomarker may be broadly applicable to patients with a variety of cancers.

According to Margaret Callahan, an investigator with the Parker Institute for Cancer Immunotherapyat MSK and the physician-researcher who led the study, the large patient cohorts, robust clinical follow-up, and rigorous statistical approach of the study gives her enthusiasm that this immune signature is telling us something important about who responds to immunotherapy and why.

The findings pave the way for prospective clinical trials designed to test whether incorporating this biomarker into patient care can improve outcomes for those who are less likely to benefit from existing therapies.

Despite the excitement surrounding immune checkpoint inhibitors, a frustrating sticking point has been the inability of doctors to predict who will benefit from them and who will not.

In making their discoveries, the researchers had data on their side. As one of the first cancer centers in the world to begin treating large numbers of patients with immunotherapy, MSK has a cache of stored blood from hundreds of patients treated over the years, efforts pioneered by MSK researchers Jedd Wolchok and Phil Wong, co-authors on the study. The investigators of this study made their discoveries using pre-treatment blood samples collected from patients enrolled on seven different clinical trials open at MSK between 2011 and 2017.

To mine the blood for clues, researchers used a technique called flow cytometry. Flow cytometry is a tool that rapidly analyzes attributes of single cells as they flow past a laser. The investigators goal was to identify markers found on patients immune cells that correlated with their response to immunotherapy primarily PD-1 targeting drugs like nivolumab (Opdivo) and pembrolizumab (Keytruda). But this wasnt a job for ordinary human eyeballs.

When you think about the fact that there are hundreds of thousands of blood cells in a single patient blood sample, and that were mapping out the composition of nearly 100 different immune cell subsets, its a real challenge to extract clinically relevant information effectively, says Ronglai Shen, a statistician in the Department of Epidemiology and Biostatistics at MSK who developed some of the statistical tools used in the study. Thats where we as data scientists were able to help Dr. Callahan and the other physician-researchers on the study. It was a perfect marriage of skills.

The statistical tools that Dr. Shen and fellow data scientist Katherine Panageas developed allowed the team to sort patients into three characteristic immune signatures, or immunotypes, based on unique patterns of blood markers.

The immunotype that jumped out was a group of patients who had high levels of a protein called LAG-3 expressed on various T cell subsets. Patients with this LAG+ immunotype, the team found, had a much shorter survival time compared with patients with a LAG- immunotype: For melanoma patients, there was a difference in median survival of more than four years (22.2 months compared with 75.8 months) and the difference was statistically significant.

This immune signature is telling us something important about who responds to immunotherapy and why.

LAG-3 (short for lymphocyte-activation gene 3) belongs to a family of molecules called immune checkpoints. Like the more well-known checkpoints CTLA-4 and PD-1, LAG-3 has an inhibitory effect on immune responses, meaning it tamps them down. Several drugs targeting LAG-3 are currently in clinical development, although defining who may benefit from them the most has been challenging.

When Dr. Callahan and her colleagues started this research, they did not plan to focus on LAG-3 specifically. We let the data lead us and LAG-3 is what shook out, she says.

One strength of the study is its use of both a discovery set and a validation set. What this means is that the investigators performed their initial analysis on one set of blood samples from a large group of patients in this case, 188 patients with melanoma. Then, they asked whether the immune signature they identified in the discovery set could predict outcomes in an entirely different batch of patients 94 people with bladder cancer.

It could, and quite well.

When we looked at our validation cohort of bladder cancer patients who received checkpoint blockade, those who had the LAG+ immunotype had a 0% response rate, Dr. Callahan says. Zero. Not one of them responded. Thats compared with a 49% response rate among people who had the LAG- immunotype.

Because of the large data set, the scientists were also able to ask how their LAG+ immunotype compares with other known biomarkers of response specifically, PD-L1 status and tumor mutation burden. What they found was the immunotype provided new and independent information about patient outcomes, rather than just echoing these other biomarkers.

Biomarkers are important in cancer for several reasons. They may help clinicians and patients select the best treatment and may allow them to avoid unnecessary treatment or treatment that is unlikely to work.

Immunotherapy drugs are not without potential toxicity, Dr. Panageas says. So, if we can spare someone the potential risks of a treatment because we know theyre not likely to respond, thats a big advance.

The second reason is cost. Immunotherapy drugs are expensive, so having a means to better match patients with available drugs is vital.

And, because the researchers identified this biomarker using patient blood samples, it raises the pleasing prospect that patients could be assessed for this marker using a simple blood draw. Other biomarkers currently in use rely on tumor tissue typically obtained by a biopsy.

If I told you that you could have a simple blood draw and in a couple of days have information to make a decision about what therapy you get, Id say it doesnt get much better than that, Dr. Callahan says. Of course, there is still much work to be done before these research findings can be applied to patients in the clinic, but we are really enthusiastic about the potential to apply these findings.

A limitation of the study is that it is retrospective, meaning that the data that were analyzed came from blood samples that were collected years ago and stored in freezers. To confirm that the findings have the potential to benefit patients, investigators will need to test their hypothesis in a prospective study, meaning one where patients are enrolled on a clinical trial specifically designed to test the idea that using this immunotype in treatment decisions can improve patient outcomes.

What Im most excited about is prospectively evaluating the idea that not only can we identify patients who wont do as well with the traditional therapies but that we can also give these patients other treatments that might help them, based on our knowledge of what LAG-3 is doing biologically, Dr. Callahan says.

Key Takeaways

Read the rest here:

MSK Study Identifies Biomarker That May Help Predict Benefits of Immunotherapy - On Cancer - Memorial Sloan Kettering

This article was originally published here

JCO Clin Cancer Inform. 2021 Aug;5:881-896. doi: 10.1200/CCI.21.00040.

ABSTRACT

Cancer Informatics for Cancer Centers (CI4CC) is a grassroots, nonprofit 501c3 organization intended to provide a focused national forum for engagement of senior cancer informatics leaders, primarily aimed at academic cancer centers anywhere in the world but with a special emphasis on the 70 National Cancer Institute-funded cancer centers. This consortium has regularly held topic-focused biannual face-to-face symposiums. These meetings are a place to review cancer informatics and data science priorities and initiatives, providing a forum for discussion of the strategic and pragmatic issues that we faced at our respective institutions and cancer centers. Here, we provide meeting highlights from the latest CI4CC Symposium, which was delayed from its original April 2020 schedule because of the COVID-19 pandemic and held virtually over three days (September 24, October 1, and October 8) in the fall of 2020. In addition to the content presented, we found that holding this event virtually once a week for 6 hours was a great way to keep the kind of deep engagement that a face-to-face meeting engenders. This is the second such publication of CI4CC Symposium highlights, the first covering the meeting that took place in Napa, California, from October 14-16, 2019. We conclude with some thoughts about using data science to learn from every child with cancer, focusing on emerging activities of the National Cancer Institutes Childhood Cancer Data Initiative.

PMID:34428097 | DOI:10.1200/CCI.21.00040

Read the original:

Cancer Informatics for Cancer Centers: Scientific Drivers for Informatics, Data Science, and Care in Pediatric, Adolescent, and Young Adult Cancer -...

The Weekend Hackathon Edition #2 The Last Hacker Standing Tea Story challenge concluded successfully on 19 August 2021. The challenge involved creating a time series analysis model that forecasts for 29 weeks . It had almost 240+participants and 110+ solutions posted on the leaderboard.

Based on the leaderboard score,we have the top 4 winners of the Tea Story Time Series Challenge, who will get free passes to the virtual Deep Learning DevCon 2021, to be held on 23-24 Sept 2021. Here, we look at the winners journeys, solution approaches and experiences at MachineHack.

First Rank Vybhav Nath C A

Vybhav Nath- a final year student at IIT Madras. He entered this field during his second year of college and started participating in MachineHack hackathons from last year. He plans to take up a career in Data Science.

Approach

He says the problem was unique in the sense that many columns in the test set had a lot of null value. So this was a challenging task to solve. He kept his preprocessing steps restricted to imputation and replacing N.S tasks. This was the first competition where he didnt use any ML model. Since many columns had null values, he interpolated the columns to get a fully populated test set. Then the final prediction was just the mean of these Price columns. He thinks this was total doosra by the cool MachineHack Team.

Experience

He says, I always participate in MH hackathons whenever possible. There are a wide variety of problems which test multiple areas. I also get to participate with many Professionals which I found to be a good pointer about where I stand among them.

Check out his solution here.

Second prize Shubham Bharadwaj

Shubham has been working as a Data Scientist for about 7 years now. He has been working on large datasets for the past 7 years. Started off with SQL then BigData Analytics, then Data Engineering and finally working as a Data Scientist. But he is new to hackathons and this is his fourth hackathon in which he has participated. He loves to solve complex problems.

Approach

The data which was provided was very raw in nature, there were around 70 percent missing values in the test dataset. From his point of view ,finding the best imputation method was the backbone of this challenge.

Preprocessing steps followed:

1. Converting the columns to correct data types,

2. Imputing the missing values- He tried various methods like filling the null values with mean of each column, mean of that row, MICE. But the best was KNN imputer with n_neighbors as 3.

For removing the outliers,he used the IQR(InterQuartile Range), which helped in reducing the mean square error.

Models tried were logistic regression, then XGBRegressor, ARIMA, T-POT, and finally H2OAutoML which yielded the best result.

Experience

Shubham says I am new to the MachineHack family, and one thing is for sure that I am here to stay. Its a great place, I have already learned so much. The datasets are of wide variety and the problem statements are unique, puzzling and complex. Its a must for every aspiring and professional data scientist to upskill themselves.

Check out his solution here.

Third prize Panshul Pant

Panshul is a Computer Science and Engineering graduate. He has picked up data science mostly from online platforms like Coursera, Hackerearth, MachineHack and by watching videos on YouTube. Going through articles on websites like Analytics India Magazine have also helped him in this journey. This problem was based on a time series which made it unique, though he solved it using machine learning algorithms rather than other traditional ways.

Approach

There were certain string values like N.S, No sale etc in all numerical columns which I changed to Null values and imputed all the null values. I tried various ways to impute NaNs like with zero, mean, f-fill and b-fill methods .Out of these forward and backward filling methods performed significantly better. Exploring the data he noticed that the prices increased over the months and years, having a trend. The target columns values were also very closely related to the average of prices of all the independent columns.He kept all data including the outliers without much change as tree based models are quite robust to outliers.

As the prices were related to time he extracted time based features as well out of which day of week proved to be useful. An average based feature which had the average of all the numerical columns was extremely useful for good predictions. He tried using some aggregate based features as well but they were not of much help. For predictions he used tree based models like lightgbm and xgboost. The combination of both of them using weighted average gave best results.

Experience

Panshul says It was definitely a valuable experience. The challenges set up by the organisers are always exciting and unique. Participating in these challenges has helped me hone my skills in this domain.

Check out his solution here.

Fourth prize Shweta Thakur

Shwetas fascination with data science started when she realised how numbers can guide decision making. She did a PGP-DSBA course from Great Learning . Even though her professional work does not involve Data Science activity, she loves to challenge herself by working on Data Science projects and participating in Hackathons.

Approach

Shweta says that the fact that it is a time series problem makes it unique. She observed the trend and seasonality in the dataset and the higher correlation between various variables. Didnt treat the outliers but tried to treat the missing values with interpolate (linear, spline)method, ffill, bfill, replacing with other values from dataset.Even though some of the features were not as significant in identifying the target but removing them didnt improve the RMSE. She tried only SARIMAX.

Experience

Shweta says It was a great experience to compete with people from different back-ground and expertise.

Check out his solution here.

Once again, join us in congratulating the winners of this exciting hackathon who indeed were the Last Hackers Standing of Tea Story- Weekend Hackathon Edition-2 . We will be back next week with the winning solutions of the ongoing challenge Soccer Fever Hackathon.

Original post:

The Winners Of Weekend Hackathon -Tea Story at MachineHack - Analytics India Magazine

Building deep and ongoing data science capabilities isn't an easy process: it takes the right people, processes and technology. Finding the right people for the right roles is an ongoing challenge, as employers and job seekers alike can attest.

"The people part is probably the least well-understood aspect of this entire equation," John Thompson, global head of advanced analytics & AI at CSL Behring, said during a virtual panel discussion on Thursday.

As the head of analytics at one of the leading international biotechnology companies, Thompson oversees data science teams that tackle a wide range of initiatives. He and the experts in the virtual panel, hosted by MLOps firm Domino Data Lab, agreed that scaling data science requires more than just data scientists.

To kick off data science initiatives at CSL Behring, Thompson says he starts with a "skeleton team you need for a project to be successful." That typically includes engineers, data scientists, a UI or UX data visualist and subject matter experts.

A successful data science team also needs a leader who can make sure projects stay focused on business objectives.

"If we're saying data science is a team sport, you don't just need all the players; you need a coach," said Matt Aslett, research director for the data, AI& Analytics Channel at 451 Research.

It's clear that a complete data science team comprises more than just data scientists -- but it isn't necessarily wise to consolidate data science teams within an IT department, added Nick Elprin, CEO and co-founder at Domino Data Lab.

"One of the things we've seen among companies we work with that are most successful is they closely align those teams with business objectives," he said. "How you guide their work and prioritize, the closer you can get that to the core company objective, the more likely you are to [be successful]. When you move more into IT, you get further away from core objectives."

Managers also need to consider how their teams are organized when they're hiring, Elprin said. They should ask, he said, "What types of skills are you going to make core to the role, and what will you augment with other people you'll collaborate with?

"Companies have success [building data science teams] with folks who know stats and basic programming and augmenting them with people who know devOps or other engineering capabilities," Elprin added.

Meanwhile, it's important to consider when professional data scientists are truly needed versus tools that purport to "democratize" data science and machine learning.

"It depends on the nature of the problem you're pointing your data science and machine learning folks toward," Elprin said. "For commoditized problems, some of the auto ML solutions can be effective. If you're talking about a problem unique to your business or core to your differentiation, you need more of... the flexibility that comes with developing your own proprietary models and using the power of code to express those ideas."

Finally, advancing impactful data science projects requires buy-in from executives, Thompson noted.

"The real challenge is the macro-level change management process; it's not really about the data science process," he said. To realize the full value of a full data science initiative, he said it's important to convey to executives that "in the end, it's going to drive change. You need to be ready to drive change... if you don't want to do that, maybe we should do a project, not a program."

Go here to see the original:

Data science is a team sport: How to choose the right players - ZDNet