Hello everyone and welcome back to Not So Random Software!
This week I have been thinking about statistics and expectations. Mental models like statistical significance and p-values help us set expectations on what are the boundaries and conditions under which we are willing to accept something as true. Without them, what would be left with?
Enjoy the random walk!
One article or paper
It is the generation of new data — the indirect discovery of unvolunteered truths about people — the makes machine learning predictions unethical? This HBR article investigates a few interesting cases.
One video or podcast
You explain to students the concept of statistical significance, and the next question often is...and now what? how do we make a decision? Statistical significance is not as trivial as looking at an arbitrary threshold. It's a model that describes our inherent uncertainty about the world.
Don't get yourself fooled by statisticians. When it comes to statistics, plenty of otherwise great scientists -- yes, even those published in peer-reviewed journals -- are doing statistics wrong. This book helps you avoid the most common traps.
One line of code
def index(conn, params) do current_user = Auth.current_user(conn) admin? = Auth.admin?(conn) page = ElixirStatus.Persistence.Posting.published(params, current_user, admin?) assigns = [ postings: page.entries, page_number: page.page_number, total_pages: page.total_pages, created_posting: load_created_posting(conn), just_signed_in: params["just_signed_in"] == "true", referred_via_elixirweekly: params["ref"] == "elixirweekly", searching?: !is_nil(params["q"]), search_query: params["q"], current_posting_filter: params["filter"] |> nil_if_empty(), posting_filters: @posting_filters, search: params["q"] |> nil_if_empty(), changeset: changeset() ] conn |> ElixirStatus.Impressionist.record("frontpage") |> render("index.html", assigns) end
You are under no obligation to remain the same person you were a year ago, a month ago, or even a day ago. You are here to create yourself, continuously.
Receive this by email
* indicates required
Email Address *