Akash Singh

I will be speaking tomorrow at the Data Architecture Sydney Conference.  My talk will focus on going beyond strategy slides when it comes to data governance.  In addition to my session, there will be other talks on topics such as data mesh, cloud data platforms, creating a single view of the customer, and AI readiness. I am excited to meet some of you there! #dataarchitecturesyd

54

2 Comments

This was an awesome Aussie TUG. Kimly Scott had tons of tips, and Heather Cox showed us things you didn't know you could do in Tableau. Sam Batchelor talked about the Tableau Social Ambassador program. If you missed it, here's your chance to play catchup. https://lnkd.in/e3cjxDN2 #Tableau #DataFam #TableauPublic #VizOfTheDay  #DataViz #TableauCommunity #TableauUsersGroup

10

1 Comment

The forgotten guiding role of data modelling... Engineers and Managers, both in my opinion, need to read and understand this. Check out the insightful article here: https://lnkd.in/gCrUbF6C.

9

🔦 Candidate Spotlight 🔦 As we steamroll through 2024 (I mean how is it May in just under 32 hours...) I'm working with a data leader on the West Coast who is looking to transition into their next organisation and help uplift the data capabilities 🚀 They've worked in the data space for nearly 2 decades across the globe and pride themselves on leading teams to implement data driven solutions that add real value to the business 📈 They have played a crucial part in many projects that have contributed to the time and money saved, and success of outcomes, including reducing costs by 85% through data platform and governance initiatives, boosting efficiency by 30% through leveraging Gen AI & seamlessly integrated a platform that allowed for a $200M+ merger to happen 💥 This person is a true gem entering the market coming with global experience and the ability to focus on best practice that suits the company to provide tangible benefits through data, so drop me a note and let me introduce you 😊 #candidatespotlight #perth #data #genAI #technology

37

1 Comment

Had an amazing time at CDAO Melbourne this week! 🎉 Although I couldn’t attend all the sessions I wanted to, here are some highlights from those I did: David Thomas offered an inspiring talk on the importance of leadership, the power of fun, and the maturity of data teams. His advice to communicate value not just through the ROI of individual use cases, but by positioning discussions within the broader context of case benefits and contributions to CDE was particularly insightful. 💡 Anshul Lyon highlighted the art of balancing data technology with data value. She emphasized that conversations with the business should be rooted in a deep understanding of their processes and clearly demonstrate how data can drive value. ⚖️ Kate Lucas delivered a compelling talk on a human-centred approach to #healthData, addressing challenges and opportunities such as funding models that consider whole-of-life care driven by linked data, data democratization that brings analytics closer to patients, and digital transformation supported by AI. Her passion for health data analytics was evident, as she reminded us through Ralph Waldo Emerson's quote: "The first wealth is health." ❤️ John Cunningham explored the challenges of #unstructuredData in achieving #GenAI success. He underscored the need to sanitize data of sensitive information before exposing it to models, along with the importance of lineage handling, deduplication, access control, and more. 🔐🤖 kudos to panellists Dr David Black, Darshil Mehta, Dr M Maruf Hossain, PhD, Dr Tim Sparks, and moderator Peter O'Donnell for a thought-provoking panel discussion on "Has AI changed what we need in a data foundation?" My key takeaways: - A solid data foundation has always been essential, and the #AIHype only reinforces its importance. 🧱 - Taking an incremental approach ensures a balance between strengthening the data foundation and delivering continuous business value. ⚙️ Great work, Corinium Global Intelligence, Vanessa Jalleh, 👏 for this thoughtfully organized event. #CDAOMel Excited to bring the learning back and share it with my team! Last but not least, I really appreciate the opportunity to chat with and thank Xi Liang in person. As an inspiring woman in data, she gave us great advice a few years back at Ada's Tribe mentor session, for which I am deeply grateful. 🤗

36

4 Comments

Stop collecting so much data. Data is everywhere, but is it all useful? Often organisations collect every possible bit of data just because of the potential they believe they can get from it. But collecting all that data comes at the cost of collecting, storing & cleansing without any guaranteed return. The key is knowing a clear use for your data, I want to collect x to be able to solve or understand y. Without that clear purpose, the data can be unused, misunderstood, or used to solve a 'problem' that doesn't exist. Three questions to ask before collecting data: 1) What is the problem you are trying to solve? 2) Who are you solving the problem for, and do they actually want it? 3) What is the overall outcome you want to achieve, what does success look like? Is there anything else you consider when collecting data? Discover more of my posts by following hashtag #datatranslator #data #analytics #datacollection

40

42 Comments

 ✍️The CDO/Head of Data job spec (need for R/Python) I chose to give the poll a 2nd chance outing but re-wrote in a non-poll fashion. Out of the list of whom a CDO/Data Head spec (with need for hands on coding skills) is likely to be drafted by the list was: ➡️CTO ➡️CEO ➡️CFO ➡️None of above ➡️Combination of above and others No. 5 appeared to be the most popular answer with No. 1 coming in 2nd and also being the highest poll choice. Further interesting responses on the topic listed below: ✅"Employers who want hands on tech exp of their CDO’s – Do they test for it at interview?" ✅"It appears of late roles such as Data Management/Data Governance/Head of Data roles have also started to list requirements for ML and AI exp? Is this becoming a trend?" ✅"A CDO will likely be more able to interact at all levels of their Data community if they have upskilled commercially via a tech route – however current hands on need for coding isn’t necessary" ✅" "hands-on" requirement in many "Head of Data" job descriptions takes away from the ultimate focus of the role, which should be driving an overall vision/strategy and guiding/mentoring your data specialists in the team to work towards" ✅“Businesses thinking that previous leadership experience is enough to be a good CDO could be one of the reasons why the average tenure of those roles is so low within the industry” ✅The poll list missed ChatGPT!!!!!! Any further thoughts? #cdojobspec #handsoncdo Edosa Odaro, Ben Steele, Dave Warder, Kate Sargent, Jon Cooke, Katy Gooblar, Mohit Joshi

3

1 Comment

Like the metaphor and have used this comparative as well, with #data being my gauges, dashboards, indicators, and VR Vision Pro heads up display. #AI is my copilot, which I trust , but always verify. My hands 🙌 on the wheel, feet on the pedals, and eyes 👀 all around are making the decisions needed at breakneck speed. Ideally, I still have a human copilot, but they would be jacked into the #GAI matrix. #datamatters #peoplemattermost #dataninjashorts from YOUR #dataninja

5

Join Emily Nota & I for the Data Market Report Webinar on Wednesday, 31st July - Online! 📊 Topics we will cover: 💡 Leadership Frustration 💡 Budget Mismatches 💡 Candidate Heavy Market 💡 Niche Roles and Internal Pressure 💡 Demand for Data Engineers 💡 Adoption of AI Gain valuable insights into the latest trends, challenges, and opportunities in the data market 📊 Perfect for those trying to navigate the market we are currently in! Link in comments to register 🔥 #DataMarketWebinar #DataEngineer #DataScience #ArtificialIntelligence #MachineLearning

18

1 Comment

Are you setting talented people in your organisation up for success? I often hear large organisations struggling with finding the right talent. And it puzzles me. These are organisations with massive data teams. Able to hire some of the best qualified graduates. Within that pool of people there has to be some exceptional talent. So what could the issues be? Organisational, Process, and Responsibility. Organisational: You hire a few exceptional people who can see how to change things and make a difference - yet they have to get their ideas signed off by committees of 20+ people. Once everyone has had their input the idea is diluted and you are 6 months down the line. Single person failure is a bigger risk in larger more regulated organisations. So your superstar needs to help bring everyone along with them, you end up limited by your best pool of people, not your best person. Process: You have to do things a certain way to fit in with legacy systems, with other departments, and with the overall organisation. Change is harder in bigger organisations, this then limits the impact talented people can achieve as they have to bring others with them. Responsibility: In a bigger organisation there are often lots of people responsible. This limits true ownership. Without that ownership people can feel like they are being held back from really making a difference. The question therefore is: Do you have a talent problem? Or is your organisational structure and process putting a ceiling on that talent? Something to ponder. Discover more of my posts by following hashtag #datatranslator #data #analytics #talent

14

4 Comments

Many Data Science and Analytics teams I observe are not effectively tracking the value they create for their organisations. Ironic, as we are supposed to be good at tracking things! With current pressures on controlling costs, this puts many Data Science teams in a precarious position. If they can’t tangibly demonstrate the value they create, then they can fall victim when organisations need to reduce headcount. In a more positive context, if they can demonstrate value then they are more likely to get the opportunity to recruit and expand. When discussing this with Data Science and Analytics teams, I often encounter the same questions. For example -  How do we measure value from the standard reports we run? How do we measure the value from innovation? These challenges can be overcome by creating an effective value measurement framework. 3 important considerations within such a framework are: 1) You need to decide what you mean by value and how you measure it. Value can be defined in a number of ways. 2) You don’t need to measure everything to demonstrate your worth. There’s nothing wrong with saying “we generated £Xm for the business, and that’s only the stuff we can track in £s” 3) You need to effectively communicate the benefits you are delivering. If no one knows about the benefits you create, then you won’t get the outcomes you want! I’m conscious a lot of Data Science and Analytics teams struggle with this. If anyone in my network would like more information about setting up and running an effective value measurement framework, then please get in touch. 

22

5 Comments

Australians are encouraged to make good choices about when and how we share our personal information… but how can we do that if companies like Meta take the choice away from us? If you’re a Facebook or Instagram user in Europe, you’ll have more autonomy than someone in Australia about whether and how Meta can use your data to create their next generation of AI products and tools. Why? Because Meta must respect Europe’s stronger privacy laws. This matters on its own terms. It’s also Exhibit #4783 of why we need the Federal Government to make good on their commitment to introduce a bill to strengthen and modernise Australia’s privacy legislation by August. Australians currently have flimsier privacy protections than people in other parts of world. This must be fixed. Great to speak with David Swan about this important issue

111

5 Comments

How To Scale Your Data Team As Your Business Grows - a tough balancing act that isn't easy to master ⚖ A lot of people in the data space have been in environments that have gone through massive growth - both from the perspective of their team and of the business. It can be a hard space to navigate as it's not a straight line and the roles you have to invest in can change depending at what stage of growth you're at. I absolutely love this article from Andy Sawyer as it dives into just that, and is a highly useful tool for leaders in the data space that are experience a period of growth - or planning one - regardless of their current stage. I'd highly recommend giving it a read as the info in there isn't something to be ignored. https://lnkd.in/eCptE9me #data #growth #hiring #scalingup

16

In today’s #AI-driven world, processes and strategies are being reimagined, which requires a new set of skills. Enter: a Head of AI. Qlik’s Brendan Grady shares the importance an AI leader to spearhead an organization’s AI efforts in @techradar. Read now: https://oal.lu/sxdOw

1

The GLAM Workbench includes a list of Australian GLAM (Galleries, Libraries, Archives & Museums) data sources with details of APIs, downloadable datasets, data portals, and Zotero compatibility. I've been updating the list over the last couple of days. More contributions are always welcome! https://lnkd.in/g3sFUVqQ There's also another long list of datasets shared by Australian GLAM organisations though government open data portals. I re-harvested all this data yesterday and it currently includes details of 22 organisations, sharing 443 datasets, containing 1,251 data files! https://lnkd.in/gvSWN3rQ You can also search the harvested metadata using Datasette-Lite: https://lnkd.in/gAV2RS-H #GLAM #HASS #openData

17

As we delve into the ever-evolving landscape of Data Science, it's clear that certain technical areas are experiencing significant growth. Here are the five biggest growth areas within the Data Science domain: 1. **Machine Learning Operations (MLOps)**: The integration of Machine Learning models into production environments is becoming increasingly sophisticated. MLOps, a practice that brings together DevOps and machine learning, is on the rise. It focuses on the automation and monitoring of all steps of machine learning system construction, including integration, testing, releasing, deployment, and infrastructure management. This ensures that models are not only accurate but also scalable and reliable in real-world applications. 2. **Automated Machine Learning (AutoML)**: The quest for efficiency has led to the rise of AutoML. This area is about automating the process of applying machine learning to real-world problems. AutoML covers the complete pipeline from the raw dataset to the deployable machine learning model. The growth in this area is driven by the need to enable data scientists to focus on more strategic tasks and to make machine learning accessible to non-experts. 3. **AI Ethics and Explainability**: As AI systems become more prevalent, the need for ethical considerations and explainability in algorithms is growing. This includes developing frameworks and tools to ensure AI systems make decisions in a fair, accountable, and transparent manner. Explainable AI (XAI) is a set of processes and methods that allows human users to comprehend and trust the results and output created by machine learning algorithms. 4. **Quantum Computing**: Although still in the nascent stages, quantum computing is set to revolutionize data science by providing unprecedented processing power. This will enable the handling of complex calculations at speeds unattainable by traditional computers, opening up new possibilities for data analysis and model complexity. 5. **Edge Computing**: With the explosion of IoT devices, edge computing has become a significant growth area in data science. Edge computing involves processing data geographically closer to where it is being generated, reducing latency and bandwidth use. This is crucial for real-time data processing tasks and is becoming increasingly important in areas such as autonomous vehicles, smart cities, and real-time analytics. These technical aspects underscore the dynamic nature of Data Science and highlight the importance of continuous learning and adaptation in the field. As these areas grow, they will undoubtedly shape the future of how data is leveraged for decision-making and innovation.

1

It's not all about the data. The data cannot tell you everything, and there are often opportunities which the data doesn't highlight. This article about Mark Cavendish, one of the best cyclists of all time, highlights this. The focus on numbers and power output is only one element of what makes him a great cyclist. The other elements are harder to measure. Timing of when to launch your sprint for the line. Knowing which wheel to follow and which gaps to go into. How to stay out of trouble when cycling at 40mph right next to other cyclists. Data is incredibly useful for pointing out easy to measure things. But don't rely on it so much that you overlook other attributes. Humans do lots of things which are hard to measure. Data can often overlook them. Discover more of my posts by following hashtag #datatranslator #data #analytics https://lnkd.in/e34xFGgE

9

4 Comments

One feature I have been looking for in #Quarto has been support for Scrollytelling. At posit::conf(2024), Andrew Bray has just presented #Closeread - a Quarto extension that enables quite a few scrollytelling techniques in Quarto. Closeread comes with different layouts and support for cross-fading plots, zooming in on code and images, and highlighted parts of a text. 🤯 Quarto is an open-source publishing system developed by Posit PBC that allows users to create and share dynamic, reproducible content like articles, presentations, dashboards, and books using #Python, #rstats, #Julia, and more, in various formats including HTML, PDF, and Word. The screengrab shows the example of Edward Tufte's description of Charles Minard's famous visualization of Napoleon's disastrous military campaign in Russia from the closeread documentation. I can't wait to try this out! #PositConf2024 #DataScience #dataviz

435

42 Comments

How much does data accuracy matter? In this case - quite a lot! While data folk bang on about data quality and accuracy, not all of it matters that much. In some cases, you just need an indication that something is working or isn't. In other cases, it is vital that your data is 100% accurate. Particularly when it deals with people's lives. In the energy industry, I have seen data that would lead to debt collectors being sent to a person's house. A person who had never been supplied by that company. Now that matters. The government knowing you are alive. That also matters. Discover more of my posts by following hashtag #datatranslator #data #analytics #dataquality

27

10 Comments