seaborn-data

seaborn-data.clj

(ns gigasquid.seaborn
  (:require [libpython-clj.require :refer [require-python]]
            [libpython-clj.python :as py :refer [py. py.. py.-]]
            [gigasquid.plot :as plot]
            [clojure.data.csv :as csv]))

(require-python '[seaborn :as sns])
(require-python '[matplotlib.pyplot :as pyplot])
(require-python '[pandas :as pandas])

;;; What is seaborn? Really cool statistical plotting

;;; sudo pip3 install seaborn

(sns/set) ;;; set default style

;;; code tutorial from https://seaborn.pydata.org/introduction.html

;;;; What is this thing that is loaded for the tidy data?
;;;; It is a dataframe
;;; Depending on each graphing function, it has the option to plot lists, numpy arrays, or DataFrames
;;; for the particular function `relplot` it requires a DataFrame

(def dots (sns/load_dataset "dots"))
(py. dots head)
;;   align  ... firing_rate
;; 0  dots  ...   33.189967
;; 1  dots  ...   31.691726
;; 2  dots  ...   34.279840
;; 3  dots  ...   32.631874
;; 4  dots  ...   35.060487

;; [5 rows x 5 columns]

(py/python-type dots) ;=>  :data-frame
(py/att-type-map dots)
(py.- dots dtypes)
;; align           object
;; choice          object
;; time             int64
;; coherence      float64
;; firing_rate    float64
;; dtype: object



;;;;;;; Now let's load the same exact data from a csv file with Clojure Data Structures

;;; we will need to covert it over to a DataFrame to finally graph it

;;; Let's try to load it from csv
(def dots-csv (slurp "https://raw.githubusercontent.com/mwaskom/seaborn-data/master/dots.csv"))

(def dots-data (doall (csv/read-csv dots-csv)))
(take 5 dots-data)
;; (["align" "choice" "time" "coherence" "firing_rate"]
;;  ["dots" "T1" "-80" "0.0" "33.1899674031"]
;;  ["dots" "T1" "-80" "3.2" "31.6917255152"]
;;  ["dots" "T1" "-80" "6.4" "34.2798399644"]
;;  ["dots" "T1" "-80" "12.8" "32.6318742985"])

;; also needs the number columns to acutally be numeric
(defn convert-numeric [row]
  (mapv #(try (Float/parseFloat %)
            (catch Exception e
              %))
      row))


(def num-dots-data (map convert-numeric dots-data))

(def dots-cols (first num-dots-data))
(def new-dots-data (rest num-dots-data))
dots-cols ;=> ["align" "choice" "time" "coherence" "firing_rate"]
(def dots-df (pandas/DataFrame new-dots-data :columns dots-cols))
(py/python-type dots-df)  ;-> :data-frame
(py.- dots-df dtypes)
;; align           object
;; choice          object
;; time           float64
;; coherence      float64
;; firing_rate    float64
;; dtype: object



(plot/with-show
  (sns/relplot :x "time" :y "firing_rate" :col "align"
               :hue "choice" :size "coherence" :style "choice"
               :facet_kws {:sharex false} :kind "line"
               :legend "full" :data dots-df))

;;;; lovely plot