3 Great Data Science Books You Can Read Now…for free

Saw this after my earlier post. Another couple of free resources.

Data Science 101

Just this week, I have become aware of 3 free online books for data science.

Data Visualization with Javascript
If you are looking for a tutorial to teach you how to make wonderful visualizations on the web, look no further. Data Visualization with JavaScript is a free online book for learning data visualization with Javascript. It provides tons of examples and step by step instructions for how to create the graphs, charts, and other visualizations. Here is a quick list of the topics:
  • Graphs
  • D3.js
  • Interactive Charts
  • Geographic Plots
  • Timelines

Frontiers in Massive Datasets

Frontiers in Massive Datasets is a report all about how science, business, communications, national security and others need to learn to handle massive amounts of data. Whether the data has been sitting in a database for years or it is now just screaming into the systems, massive data is now a problem for almost every industry…

View original post 148 more words

A Big Data How To Guide – Overlooked advice?

Lots of thoughtful guidance for engineers, designers, scientists and businesses to focus on while doing the exciting interdisciplinary and real work that lies ahead.

The extremely thorough study made a few major recommendations, from my initial read:

  1.  Deeply interdisciplinary engineering approach influenced by design and system thinking
  2. “Middleware” to link high level analysis with distributed computational resources that maximizes reuse [Consider parallel computing reference by Asanovic and team]
  3. Students and the workforce need foundational statistical and computational training [Think ubiquitous driver permits after mass adoption of automobiles]

They allude to the possibility of an FFT like standard tool for massive data analysis (anyone who has done signal processing knows how ubiquitous the FFT is in getting almost anything done more efficiently). However, they claim to be pessimistic given the scope/variety of this domain.

Perhaps these principles can be uncovered once and for all, such that each
successive generation of researchers does not need to reconsider the massive data problem afresh.

Will we create a platform that stands for decades (think Shannon and Schockley) or take short cuts that lead to yet another disappointing bubble of frenetic activity?

Time will tell – Metonymy Labs seeks to participate productively.

Startup culture

Screen Shot 2014-10-22 at 9.41.52 PM

There are risks in creating a “process culture” or to try and institute top-down initiatives in an early venture. Reality in a startup is just too volatile to try and plan for. Starting a venture or being part of a cofounding team is not like any other job. Many of us (me especially!) who have spent time in the corporate world forget this.

One of the things that seems to be helpful is to develop a “warrior for a cause mindset” in building a team. Spending the time up front in introspecting why this is an important venture for each team member to pursue. Without a team member being fanatically action oriented and driven towards realizing the cause no talent/skill likely matter. Job titles will definitely evolve as the venture progresses – action and the cause seem to be the only constants.

In the first several months and beyond it seems to me there are really two things that add value and each member gets to measure contributions in one of two dimensions –
1. Make product
2. Sell users/customers

If we don’t see on a daily/weekly basis how we are individually creating progress in either of these then we should question what we are doing – it likely doesn’t matter.

There is an underlying existential aspect of resource commitment to the cause (funding, team building, etc.) but this will follow from successes in 1 & 2.

[incorporating other’s thoughts I am sure – have read too many posts to be able to decipher provenance without lots of effort]