3  Visualize Data

(ns assignment.visualize-data
  (:require
    [aerial.hanami.templates :as ht]
    [assignment.generate-data :refer [data]]
    [scicloj.kindly.v4.kind :as kind]
    [scicloj.ml.dataset :as ds]
    [scicloj.noj.v1.vis.hanami :as hanami]))

(-> data
    ds/shuffle
    (ds/head 7))

_unnamed [7 3]:

:x1	:x2	:group
1.15632892	-2.34357298	normal
3.67532450	5.10155574	normal
5.40007373	2.62960971	gamma
9.36665576	6.72907653	gamma
1.59971439	-2.67785685	normal
8.55412024	4.42748644	gamma
8.08554275	11.39241993	gamma

(def cols-of-interest [:x1 :x2])

3.1 Group: normal

(def norm-dat
  (-> data
      (ds/select-rows #(= (:group %) "normal"))))

^kind/vega
(let [dat (ds/rows norm-dat :as-maps)
      column-names cols-of-interest]
  {:data   {:values dat}
   :repeat {:column column-names}
   :spec   {:mark     "bar"
            :encoding {:x     {:field {:repeat "column"}
                               :bin   {:steps [1 3]} :type "quantitative"}
                       :y     {:aggregate "count"}
                       :color {:field :group}}}})

(ds/info norm-dat)

_unnamed: descriptive-stats [3 12]:

:col-name	:datatype	:n-valid	:n-missing	:min	:mean	:mode	:max	:standard-deviation	:skew	:first	:last
:x1	:float64	200	0	-3.52032814	3.15359995		8.96857501	1.98163550	-0.22145647	3.581	5.726
:x2	:float64	200	0	-12.27623596	-0.44328147		13.93989001	4.29960213	0.13687752	1.448	-3.545
:group	:string	200	0			normal				normal	normal

3.2 Group: gamma

(def gamma-dat
  (ds/select-rows data #(= (:group %) "gamma")))

^kind/vega
(let [dat (ds/rows gamma-dat :as-maps)
      column-names cols-of-interest]
  {:data   {:values dat}
   :repeat {:column column-names}
   :spec   {:mark     "bar"
            :encoding {:x     {:field {:repeat "column"}
                               :bin   {:steps [1 3]} :type "quantitative"}
                       :y     {:aggregate "count"}
                       :color {:field :group}}}})

(ds/info gamma-dat)

_unnamed: descriptive-stats [3 12]:

:col-name	:datatype	:n-valid	:n-missing	:min	:mean	:mode	:max	:standard-deviation	:skew	:first	:last
:x1	:float64	200	0	2.46503412	7.06417638		12.26392965	1.85266087	0.24642603	2.465	5.841
:x2	:float64	200	0	1.95034441	7.50807335		19.84870098	3.48813161	0.84906043	8.546	19.85
:group	:string	200	0			gamma				gamma	gamma

3.3 Group: log-normal

(def log-normal-dat
  (ds/select-rows data #(= (:group %) "log-normal")))

^kind/vega
(let [dat (ds/rows log-normal-dat :as-maps)
      column-names cols-of-interest]
  {:data   {:values dat}
   :repeat {:column column-names}
   :spec   {:mark     "bar"
            :encoding {:x     {:field {:repeat "column"}
                               :bin   {:steps [1 3]} :type "quantitative"}
                       :y     {:aggregate "count"}
                       :color {:field :group}}}})

(ds/info log-normal-dat)

_unnamed: descriptive-stats [3 12]:

:col-name	:datatype	:n-valid	:n-missing	:min	:mean	:mode	:max	:standard-deviation	:skew	:first	:last
:x1	:float64	200	0	-5.96223122	-0.98826246		4.59510840	1.99868266	0.16848382	-3.653	-3.305
:x2	:float64	200	0	-30.50059718	-6.18813429		-1.58353073	3.23981886	-2.81709917	-5.231	-4.002
:group	:string	200	0			log-normal				log-normal	log-normal

3.4 Full data

(-> data
    (hanami/plot ht/point-chart
                 {:X "x1" :Y "x2" :COLOR "group"}))

^kind/vega
(let [dat (ds/rows data :as-maps)
      column-names cols-of-interest]
  {:data   {:values dat}
   :repeat {:column column-names}
   :spec   {:mark     "bar"
            :encoding {:x     {:field {:repeat "column"}
                               :bin   {:steps [1 3]} :type "quantitative"}
                       :y     {:aggregate "count"}
                       :color {:field :group}}}})

(comment
  (defn dist-range [dist]
    (-> (apply max dist)
        (-
          (apply min dist)))))

(comment                                                    ; live works, but wont render
  (hanami/hconcat norm-dat {}
                  [(hanami/histogram norm-dat :x1 {:nbins 20})
                   (hanami/histogram norm-dat :x2 {:nbins 20})])

  (hanami/hconcat gamma-dat {}
                  [(hanami/histogram gamma-dat :x1 {:nbins 20})
                   (hanami/histogram gamma-dat :x2 {:nbins 20})])

  (hanami/hconcat log-normal-dat {}
                  [(hanami/histogram log-normal-dat :x1 {:nbins 20})
                   (hanami/histogram log-normal-dat :x2 {:nbins 20})])

  ^kind/vega
  (let [data (ds/rows data :as-maps)]
    {:data     {:values data}
     :mark     "bar"
     :encoding {:x     {:field :x1 :bin {:steps [2 3]} :type "quantitative"}
                :y     {:aggregate "count"}
                :color {:field :group}}}))

source: src/assignment/visualize_data.clj