Cloud Hosting

Data scientist is one of the hottest jobs in IT. Companies are increasingly eager to hire data professionals who can make sense of the wide array of data the business collects. If you are looking to get into this lucrative field, or want to stand out against the competition, certification can be key.

Data science certifications give you an opportunity not only to develop skills that are hard to find in your desired industry but also to validate your data science know-how so that recruiters and hiring managers know what theyre getting if they hire you.

Whether youre looking to earn a certification from an accredited university, gain experience as a new grad, hone vendor-specific skills, or demonstrate your knowledge of data analytics, the following certifications (presented in alphabetical order) will work for you.

The Certified Analytics Professional (CAP) is a vendor-neutral certification that validates your ability to transform complex data into valuable insights and actions, which is exactly what businesses are looking for in a data scientist: someone who understands data, can draw logical conclusions and express to key stakeholders why those data points are significant. Youll need to apply and meet certain criteria before you can take the CAP or the associate level aCAP exams. To qualify for the CAP certification exam, youll need three years of related experience if you have a masters degree in a related field, five years of related experience if you hold a bachelors in a related field, or seven years of experience if you have any degree unrelated to analytics. To qualify for the aCAP exam, you will need a masters degree and less than three years of related experience in data or analytics.

Cost: CAP exam: $495 for INFORMS members, $695 for non-members; aCAP exam: $200 for INFORMS members, $300 for non-members

Location: In person at designated test centers

Duration: Self-paced

Expiration: Valid for three years

Cloudera has discontinued its Cloudera Certified Professional (CCP) and Cloudera Certified Associate (CCA) certifications in favor of the new Cloudera Data Platform (CDP) Generalist certification, which verifies proficiency with the platform. The new exam tests general knowledge of the platform and applies to multiple roles, including administrator, developer, data analyst, data engineer, data scientist, and system architect. The exam consists of 60 questions and the candidate has 90 minutes to complete it.

Cost: $300

Location: Online

Duration: 90 minutes

Expiration: Valid for two years

The Data Science Council of America (DASCA) Senior Data Scientist (SDS) certification program is designed for professionals with five or more years of experience in research and analytics. Its recommended that students have knowledge of databases, spreadsheets, statistical analytics, SPSS/SAS, R, quantitative methods, and the fundamentals of object-oriented programming and RDBMS. The program includes five tracks that will appeal to a range of candidates each track has differing requirements in terms of degree-level, work experience, and prerequisites to apply. Youll need at least a bachelors degree and more than five years of experience in data science to be eligible for each track, while other tracks require a masters degree or past certifications.

Cost: $775

Location: Online

Duration: Self-paced

Expiration: 5 years

The Data Science Council of America (DASCA) offers the Principal Data Scientist (PDS) certification, which includes three tracks for data science professionals with 10 or more years of experience in big data. The exam covers everything from fundamental to advanced data science concepts such as big data best practices, business strategies for data, building cross-organizational support, machine learning, natural language processing, scholastic modeling, and more. The exam is designed for seasoned and high-achiever data science thought and practice leaders.

Cost: $850, track 1; $1,050, track 2; $750, Track 3; $1,250, track 4

Location: Online

Duration: Self-paced

Expiration: Credentials do not expire

The IBM Data Science Professional Certificate consists of nine courses on data science, open source tools, data science methodology, Python, Databases and SQL, data analysis, data visualization, machine learning, and a final applied data science capstone. The certification coursework takes place online through Coursera with a flexible schedule and takes an average of three months to complete, but you are free to take more or less time. The course includes hands-on projects to help you build a portfolio to showcase your data science talents to potential employers.

Cost: Free

Location: Online

Duration: Self-paced

Expiration: Credentials do not expire

Microsofts Azure AI Fundamentals certification validates your knowledge of machine learning and artificial intelligence concepts and how they relate to Microsoft Azure services. Its a fundamentals exam, so you dont need extensive experience to pass the exam. Its a good place to start if you are new to AI or AI on Azure and want to demonstrate your skills and knowledge to employers.

Cost: $99

Location: Online

Duration: Self-paced

Expiration: Credentials do not expire

The Azure Data Scientist Associate certification from Microsoft focuses your ability to utilize machine learning to implement and run machine learning workloads on Azure. Candidates for the exam are tested on ML, AI solutions, NLP, computer vision, and predictive analytics. You will need to be skilled in deploying and managing resources, managing identities and governance, implementing and managing storage, and configuring and managing virtual networks.

Cost: $165

Location: Online

Duration: Self-paced

Expiration: Credentials do not expire

The Open Group Professional Certification Program for the Data Scientist Professional (Open CDS) is an experience-based certification without any traditional training courses or exams. Youll start at level one as a Certified Data Scientist, then you can move to level two where youll become a Master Certified Data Scientist and finally, you can pass the third level to become a Distinguished Certified Data Scientist. Certification requires a three-step process that includes applying for the certification, completing the experience application form, and attending a board review.

Cost: Contact for pricing

Location: On-site

Duration: Varies by level

Expiration: Credentials do not expire

The AI and Machine Learning Professional certification from SAS demonstrates your ability to use open source tools to gain insight from data using AI and analytics skills. The certification consists of several exams that cover topics such as machine learning, natural language processing, computer vision, and model forecasting and optimization. Youll need to pass the SAS Certified Specialist exams in Machine Learning, Forecasting and Optimization, and Natural Language Processing and Computer Vision to earn the AI and Machine Learning professional designation.

Cost: $180 per exam

Location: Online

Duration: Self-paced

Expiration: Credentials do not expire

The SAS Certified Advanced Analytics Professional Using SAS 9 credential validates your ability to analyze big data with a variety of statistical analysis and predictive modeling techniques. Youll need experience in machine learning and predictive modeling techniques, including their use with big, distributed, and in-memory data sets. You should also have experience with pattern detection, experimentation in business optimization techniques, and time-series forecasting. The certification requires passing three exams: Predictive Modeling Using SAS Enterprise Miner 7, 13, or 14; SAS Advanced Predictive Modeling; and SAS Text Analytics, Time Series, Experimentation, and Optimization.

Cost: $250 for the Predictive Modeling Using SAS Enterprise Miner exam; $180 each for the other two required exams

Location: Online

Duration: Self-paced

Expiration: Credentials do not expire

The SAS Certified Data Scientist certification is a combination of the other two data certifications offered through SAS. It covers programming skills; managing and improving data; transforming, accessing, and manipulating data; and how to work with popular data visualization tools. Once you earn both the Big Data Professional and Advance Analytics Professional certifications, you can qualify to earn your SAS Certified Data Scientist designation. Youll need to complete all 18 courses and pass the five exams between the two separate certifications.

Cost: $180 per exam

Location: Online

Duration: Self-paced

Expiration: Credentials do not expire

The TensorFlow Developer Certificate program is a foundational certificate for students, developers, and data scientists who want to demonstrate practical machine learning skills through the building and training of models using TensorFlow. The exam tests your knowledge of and ability to integrate machine learning into various tools and applications. To pass the exam you will need to be experienced with the foundational principles of ML and deep learning, building ML models, image recognition algorithms, deep neural networks, and natural language processing.

Cost: $100 per exam

Location: Online

Duration: Self-paced

Expiration: Credentials do not expire

More on data science:

Follow this link:

15 top data science certifications | CIO

Read More..

Analysts estimate that by 2025, 30% of generated data will be real-time data. That is 52 zettabytes (ZB) of real-time data per year roughly the amount of total data produced in 2020. Since data volumes have grown so rapidly, 52 ZB is three times the amount of total data produced in 2015. With this exponential growth, its clear that conquering real-time data is the future of data science.

Over the last decade, technologies have been developed by the likes of Materialize, Deephaven, Kafka and Redpanda to work with these streams of real-time data. They can transform, transmit and persist data streams on-the-fly and provide the basic building blocks needed to construct applications for the new real-time reality. But to really make such enormous volumes of data useful, artificial intelligence (AI) must be employed.

Enterprises need insightful technology that can create knowledge and understanding with minimal human intervention to keep up with the tidal wave of real-time data. Putting this idea of applying AI algorithms to real-time data into practice is still in its infancy, though. Specialized hedge funds and big-name AI players like Google and Facebook make use of real-time AI, but few others have waded into these waters.

To make real-time AI ubiquitous, supporting software must be developed. This software needs to provide:

Developers and data scientists want to spend their time thinking about important AI problems, not worrying about time-consuming data plumbing. A data scientist should not care if data is a static table from Pandas or a dynamic table from Kafka. Both are tables and should be treated the same way. Unfortunately, most current generation systems treat static and dynamic data differently. The data is obtained in different ways, queried in different ways, and used in different ways. This makes transitions from research to production expensive and labor-intensive.

To really get value out of real-time AI, developers and data scientists need to be able to seamlessly transition between using static data and dynamic data within the same software environment. This requires common APIs and a framework that can process both static and real-time data in a UX-consistent way.

The sexiest work for AI engineers and data scientists is creating new models. Unfortunately, the bulk of an AI engineers or data scientists time is devoted to being a data janitor. Datasets are inevitably dirty and must be cleaned and massaged into the right form. This is thankless and time-consuming work. With an exponentially growing flood of real-time data, this whole process must take less human labor and must work on both static and streaming data.

In practice, easy data cleaning is accomplished by having a concise, powerful, and expressive way to perform common data cleaning operations that works on both static and dynamic data. This includes removing bad data, filling missing values, joining multiple data sources, and transforming data formats.

Currently, there are a few technologies that allow users to implement data cleaning and manipulation logic just once and use it for both static and real-time data. Materialize and ksqlDb both allow SQL queries of Kafka streams. These options are good choices for use cases with relatively simple logic or for SQL developers. Deephaven has a table-oriented query language that supports Kafka, Parquet, CSV, and other common data formats. This kind of query language is suited for more complex and more mathematical logic, or for Python developers.

Many possibly even most new AI models never make it from research to production. This hold up is because research and production are typically implemented using very different software environments. Research environments are geared towards working with large static datasets, model calibration, and model validation. On the other hand, production environments make predictions on new events as they come in. To increase the fraction of AI models that impact the world, the steps for moving from research to production must be extremely easy.

Consider an ideal scenario: First, static and real-time data would be accessed and manipulated through the same API. This provides a consistent platform to build applications using static and/or real-time data. Second, data cleaning and manipulation logic would be implemented once for use in both static research and dynamic production cases. Duplicating this logic is expensive and increases the odds that research and production differ in unexpected and consequential ways. Third, AI models would be easy to serialize and deserialize. This allows production models to be switched out simply by changing a file path or URL. Finally, the system would make it easy to monitor in real time how well production AI models are performing in the wild.

Change is inevitable, especially when working with dynamic data. In data systems, these changes can be in input data sources, requirements, team members and more. No matter how carefully a project is planned, it will be forced to adapt over time. Often these adaptations never happen. Accumulated technical debt and knowledge lost through staffing changes kill these efforts.

To handle a changing world, real-time AI infrastructure must make all phases of a project (from training to validation to production) understandable and modifiable by a very small team. And not just the original team it was built for it should be understandable and modifiable by new individuals that inherit existing production applications.

As the tidal wave of real-time data strikes, we will see significant innovations in real-time AI. Real-time AI will move beyond the Googles and Facebooks of the world and into the toolkit of all AI engineers. We will get better answers, faster, and with less work. Engineers and data scientists will be able to spend more of their time focusing on interesting and important real-time solutions. Businesses will get higher-quality, timely answers from fewer employees, reducing the challenges of hiring AI talent.

When we have software tools that facilitate these four requirements, we will finally be able to get real-time AI right.

Chip Kent is the chief data scientist at Deephaven Data Labs.

Welcome to the VentureBeat community!

DataDecisionMakers is where experts, including the technical people doing data work, can share data-related insights and innovation.

If you want to read about cutting-edge ideas and up-to-date information, best practices, and the future of data and data tech, join us at DataDecisionMakers.

You might even considercontributing an articleof your own!

Read More From DataDecisionMakers

Read the rest here:

Roadblocks to getting real-time AI right - VentureBeat

Read More..

Long before Leah Boustan was a professor of economics at Princeton, she was a Princeton undergraduate putting the final touches on her senior thesis.

Working alongside her advisor, longtime professor Henry Hank Farber, Boustan published a 100-page research project that compared outcomes for students who dropped out of high school in the early 1960s with those who dropped out decades later.

"I cant remember the exact moment I decided to become an economic historian, but I remember telling Hank I was really interested in comparisons of cohorts over time," Boustan said. "That interest is the basis for a lot of my work even today."

Twenty-two years after graduating from Princeton, Boustan has published a book that uses troves of data and the latest innovations in data science to examine an issue Boustan considers "one of the most fraught issues in U.S. politics" both today and in the past: immigration.

Leah Boustan as a Princeton undergraduate with her father, Harlan Platt

Photo courtesy of Leah Boustan

Written with her longtime collaborator Ran Abramitzky of Stanford University, Streets of Gold: America's Untold Story of Immigrant Success introduces the public to more than a decade of her rigorous, empirical research on the personal and society-wide impacts of immigration.

Weaving together personal family stories including their own with insights from the data, Boustan and Abramitzky tell an uplifting story about the promise of immigration. One finding Boustan found particularly surprising is how well children of immigrants have done economically, both today and in the past.

The fact that children of immigrants who came from poor families in the 1980s moved up the economic ladder at the same pace as children of the Ellis Island generation that floored me," said Boustan.

One hundred years ago, Italy a major sending country of immigrants to the U.S. had about half the GDP per capita of the United States. Once in America, however, the sons of Italian immigrants rose up. Those who grew up in the 25th percentile of income distribution in the late 1800s earned enough as adults to be near the 60th percentile.

Today, children of immigrants from Nicaragua, which has about one-tenth the GDP per capita of the United States, see similar rates of economic mobility.

"Theres no reason that has to be true but it turned out to be," Boustan said. "It's something really remarkable we're able to see because of the data."

That data and the methodologies Boustan and Abramitzky developed to make use of it deserve almost as much attention as the findings.

In addition to working with and linking modern data like IRS tax records and birth certificate files, a partnership with the genealogy website Ancestry.com made it possible for Boustan and Abramitzky to automate searches and follow millions of families over more than 100 years of Census data. From there, they worked with audio recordings of historical interviews and congressional speeches, using machine-learning tools to analyze these texts and glean big-data insights.

The rigor of the research is one reason it's so groundbreaking, and a tradition Boustan can trace to her days as a Princeton undergraduate.

As a high school debate student interested in public policy, Boustan applied early decision to Princeton with the aim of declaring a concentration at the School of Public and International Affairs (SPIA). But her time learning from Professor Farber changed her mind.

"I took economics classes in order to major in SPIA, she said. One of those classes was ECO 313: Applied Econometrics with Hank Farber, where we used real data sets to answer questions. I fell in love with that class."

Boustan told Farber she wanted to spend a summer working in Washington, but he persuaded her to stay in Princeton instead to learn more about data analysis and to see how building expertise in a discipline like economics could help her produce the kind of policy-relevant work that legislators really need.

"So thats what I did, and I never really looked back," she said.

Boustan declared Economics as her concentration at Princeton and started spending her free time in the computer lab of the Industrial Relations (IR) Section a group widely known for training and supporting some of the most famous labor economists and empiricists in the field, including 2021 Nobel Laureates and Princeton alumni David Card and Joshua Angrist.

From there, she went on to earn a Ph.D. from Harvard. After several years teaching at the University of California Los Angeles, Boustan returned to Princeton in 2017.

As Boustan hits the road to talk about her new book, she's able to marvel at how things have come full circle. When Farber signed on as Boustan's undergraduate senior thesis advisor, he was the director of the IR Section. Last year, Boustan herself was given the title, an honor she doesn't take for granted.

The IR Section is a true intellectual community, she said. The faculty sit right beside the graduate students almost like in a lab and work closely together. And the research coming out of the Section is always connected to the real world, from minimum wage to unemployment to the immigration work that I have been doing.

For Farber, who says Boustan was one of the best undergraduate students he had the pleasure of teaching in his 30 years at Princeton, having Boustan as a colleague has been a source of some pride.

Farber also noted how Boustan, no longer the student, has herself excelled in role of advisor. Leah has really played a key role in guiding the IR Section not only on the research side, but on the teaching side as well, he said.

In addition to committing her time as an advisor to dozens of undergraduate and graduate students, Boustan recently took on the task of teaching Princetons Principles of Microeconomics course a popular class for undergraduates across a wide range of majors.

I taught that class for many years myself, said Farber. It was wonderful to see Leah make it her own and take it in a whole new direction. The reaction from students this year was very positive.

Boustan says her research and her role as an economic historian give her hope for the future of immigration policy.

Sometimes we feel so stuck. We feel polarized. Congress cant pass legislation. On immigration, weve been at a stalemate for 50 years. But you look at history, and you see weve had wild change. It reminds me Im living in one small moment in history.I think economic history helps us recognize the possibility of scope for change.

Specifically, she hopes much-needed policy change will come for the Dreamers, undocumented immigrants who came to the United States as children. "Our research shows the most optimistic vision of what the children of immigrants can achieve," she said.

Because it can take 30 to 40 years to follow children into the labor market, her research on modern immigration focuses on children born in the 1980s. These children lived in households who benefited from immigration amnesty programs during the Reagan administration. Boustan worries that studies of more recent immigrant arrivals many of whom are undocumented without any path to citizenship could produce less optimistic findings.

"Im worried about the next generation and what I'll find when we write Streets of Gold 2.0," she said. "Theres lots of promise for children of immigrants if they and their parents have some pathway into the formal labor market. I think its urgent to pass DACA as legislation and really return to the idea of comprehensive immigration reform."

Readers interested in learning more can read about five immigration myths dispelled in "Streets of Gold." This Thursday, June 9, at 8:30 p.m. ET, Professor Boustan will answer questions about her research in a Twitter Spaces event with Joey Politano, author of the Apricitas Economics blog. Join the event.

Read the original post:

Modern immigrants' children have climbed the economic ladder as fast as the Ellis Island generation - Princeton University

Read More..

Regression models, both single and multivariate, are the backbone of many types of machine learning. Using the structure you specify, these tools create equations that match the modeled data set as closely as possible. Regression algorithms create the optimal equation by minimizing the error between the results predicted by the model and the provided data.

That said, no regression model will ever be perfect (and if your model does appear to be nearly perfect I recommend you check for overfitting). There will always be a difference between the values predicted by a regression model and the actual data. Those differences will change dramatically as you change the structure of the model, which is where residuals come into play.

The residual for a specific data point is the difference between the value predicted by the regression and the observed value for that data point. Calculating the residual provides a valuable clue into how well your model fits the data set. To calculate residuals we need to find the difference between the calculated value for the independent variable and the observed value for the independent variable.

The residual for a specific data point is the difference between the value predicted by the regression and the observed value for that data point. Calculating the residual provides a valuable clue into how well your model fits the data set. A poorly fit regression model will yield residuals for some data points that are very large, which indicates the model is not capturing a trend in the data set. A well-fit regression model will yield small residuals for all data points.

Lets talk about how to calculate residuals.

In order to calculate residuals we first need a data set for the example. We can create a fairly trivial data set using Pythons Pandas, NumPy and scikit-learn packages. You can use the following code to create a data set thats essentially y = x with some noise added to each point.

That code performs the following steps:

We can now use that data frame as our sample data set.

Want More Data Science Tutorials? We Got You.How to Use Float in Python (With Sample Code!)

The Dependent variable is our x data series, and the Independent variable is our y. Now we need a model which predicts y as a function of x. We can do that using scikit-learns linear regression model with the following code.

That code works as follows:

If the model perfectly matches the data set, then the values in the Calculated column will match the values in the Dependent column. We can plot the data to see if it does or not.

nope.

We could have seen that coming because we used a first-order linear regression model to match a data set with known noise in it. In other words, we know that this model would have perfectly fit y = x, but the variation we added in each data point made everyy a bit different from the corresponding x. Instead of perfection, we see gaps between the Regression line and the Data points. Those gaps are called the residuals. See the following plot which highlights the residual for the point at x = 4.

To calculate the residuals we need to find the difference between the calculated value for the independent variable and the observed value for the independent variable. In other words, we need to calculate the difference between the Calculated and Independent columns in our data frame. We can do so with the following code:

We can now plot the residuals to see how they vary across the data set. Heres an example of a plotted output:

Notice how some of the residuals are greater than zero and others are less than zero. This will always be the case! Since linear regression reduces the total error between the data and the model to zero the result must contain some errors less than zero to balance out the errors that are greater than zero.

You can also see how some of the errors are larger than others. Several of the residuals are in the +0.25 0.5 range, while others have an absolute value in the range of 0.75 1. These are the signs that you look for to ensure a model is well fit to the data. If theres a dramatic difference, such as a single point or a clustered group of points with a much larger residual, you know that your model has an issue. For example, if the residual at x = 4 was -5 that would be a clear sign of an issue. Note that a residual that large would probably indicate an outlier in the data set, and you should consider removing the point using interquartile range (IQR) methods.

More on IQR? Coming Right Up.How to Find Outliers With IQR Using Python

To highlight the argument that residuals can demonstrate a poor model fit, lets consider a second data set. To create the new data set I made two changes. The changed lines of code are as follows:

The first change increased the length of the data frame index to 100. This created a data set with 100 points, instead of the prior 10. The second change made the Dependent variable be a function of the Independent variable squared, creating a parabolic data set. Performing the same linear regression as before (not a single letter of code changed) and plotting the data presents the following:

Since this is just an example meant to demonstrate the point, we can already tell that the regression doesnt fit the data well. Theres an obvious curve to the data, but the regression is a single straight line. The regression under-predicts at the low and high ends, and over-predicts in the middle. We also know its going to be a poor fit because its a first-order linear regression on a parabolic data set.

That said, this visualization effectively demonstrates how examining the residuals can show a model with a poor fit. Consider the following plot, which I generated using the exact same code as the prior residuals plot.

Can you see the trend in the residuals? The residuals are very negative when the X Data is both low and high, indicating that the model is under-predicting the data at those points. The residuals are also positive when the X Data is around the midpoint, indicating the model is over-predicting the data in that range. Clearly the model is the wrong shape and, since the residuals curve only shows one inflection point, we can reasonably guess that we need to increase the order of the model by one (to two).

If we repeat the process using a second-order regression, we obtain the following residuals plot.

The only discernible pattern here is that the residuals increase as the X Data increases. Since the Dependent data includes noise, which is a function of the X Data, we expect that to happen. What we dont see are very large residuals, or indications of a different shape to the data set. This means we now have a model that fits the data set well.

And with that, youre all set to start evaluating the performance of your machine learning models by calculating and plotting residuals!

See the article here:

How to Find Residuals in Regression Analysis - Built In

Read More..