EDA update

marcos-mansur · Oct 16, 2021 · 721a210 · 721a210
1 parent ea7fe7d
commit 721a210
Showing 1 changed file with 72 additions and 33 deletions.
diff --git a/Notebook/EDA.ipynb b/Notebook/EDA.ipynb
@@ -326,7 +326,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": 4,
    "id": "0e045ad6",
    "metadata": {},
    "outputs": [],
@@ -422,7 +422,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 8,
+   "execution_count": 5,
    "id": "df3d4c9f",
    "metadata": {},
    "outputs": [],
@@ -448,7 +448,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 10,
+   "execution_count": 6,
    "id": "5beb5971",
    "metadata": {},
    "outputs": [
@@ -497,7 +497,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 135,
+   "execution_count": 7,
    "id": "15cd2efc",
    "metadata": {},
    "outputs": [],
@@ -523,9 +523,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 136,
+   "execution_count": 19,
    "id": "bd2362f4",
-   "metadata": {},
+   "metadata": {
+    "scrolled": true
+   },
    "outputs": [
     {
      "data": {
@@ -652,20 +654,22 @@
        "4  2020-1  "
       ]
      },
-     "execution_count": 136,
+     "execution_count": 19,
      "metadata": {},
      "output_type": "execute_result"
     }
    ],
    "source": [
     "# cria um dataset aplicando a transformação\n",
     "df_time = get_time_features.transform(df_type)\n",
+    "df_time.loc[df_time['DT_VENDA']>='2021-1-1','MES'] = 13\n",
+    "df_time.loc[df_time['DT_VENDA']>='2021-2-1','MES'] = 14\n",
     "df_time.head()"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 54,
+   "execution_count": 20,
    "id": "6e0a244e",
    "metadata": {},
    "outputs": [
@@ -710,7 +714,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "79ecca08",
+   "id": "fc44fa49",
    "metadata": {},
    "source": [
     "### Valor total por dia"
@@ -757,26 +761,15 @@
   },
   {
    "cell_type": "markdown",
-   "id": "f68edad7",
+   "id": "89f15828",
    "metadata": {},
    "source": [
     "### Valor mensal"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 146,
-   "id": "8a93e97c",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "df_time.loc[df_time['DT_VENDA']>='2021-1-1','MES'] = 13\n",
-    "df_time.loc[df_time['DT_VENDA']>='2021-2-1','MES'] = 14"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 155,
+   "execution_count": 11,
    "id": "b34d60ce",
    "metadata": {},
    "outputs": [],
@@ -789,7 +782,7 @@
   {
    "cell_type": "code",
    "execution_count": 233,
-   "id": "6db76ce9",
+   "id": "60b2028b",
    "metadata": {},
    "outputs": [
     {
@@ -820,7 +813,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "b5087188",
+   "id": "c8536dbd",
    "metadata": {},
    "source": [
     "Os valores de compra tem desvio padrão maior que a própria média, isso é um spread significativo. O desvio padrão aparenta seguir a mesma tendência da média."
@@ -829,7 +822,7 @@
   {
    "cell_type": "code",
    "execution_count": 245,
-   "id": "07a3388a",
+   "id": "8952fc6c",
    "metadata": {},
    "outputs": [
     {
@@ -860,15 +853,15 @@
   },
   {
    "cell_type": "markdown",
-   "id": "7a1e1350",
+   "id": "8229b5ad",
    "metadata": {},
    "source": [
     "O pico em 2020-12 ocorre tanto no valor total quanto na média, entretanto, o pico é mais acentuado quando olhamos o valor vendido total. Isso sugere que a quantidade de compras cresceu muito em dezembro. Vamos checar essa relação."
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "c905d9fd",
+   "id": "0253e38a",
    "metadata": {},
    "source": [
     "### Número de Compras"
@@ -877,7 +870,7 @@
   {
    "cell_type": "code",
    "execution_count": 249,
-   "id": "ce7b76e8",
+   "id": "ccce38ba",
    "metadata": {},
    "outputs": [
     {
@@ -903,7 +896,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "ab09c357",
+   "id": "3ea571c2",
    "metadata": {},
    "source": [
     "## Valor por canal"
@@ -912,7 +905,7 @@
   {
    "cell_type": "code",
    "execution_count": 197,
-   "id": "ea598b9d",
+   "id": "90b69086",
    "metadata": {},
    "outputs": [
     {
@@ -947,7 +940,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "79ef0ca1",
+   "id": "79f01d9f",
    "metadata": {},
    "source": [
     "### Dividindo os anos"
@@ -956,7 +949,7 @@
   {
    "cell_type": "code",
    "execution_count": 251,
-   "id": "5f8985e6",
+   "id": "8b4d3509",
    "metadata": {},
    "outputs": [
     {
@@ -1292,10 +1285,56 @@
     "test_stationarity(df_test)"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "id": "5748c5f1",
+   "metadata": {},
+   "source": [
+    "### Número de dias por mês\n",
+    "\n",
+    "Vamos ver quantos dias válidos temos em cada mês."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "id": "0e9de500",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "MES  ANO+MES\n",
+       "1    2020-1     30\n",
+       "2    2020-2     28\n",
+       "3    2020-3     29\n",
+       "4    2020-4     26\n",
+       "5    2020-5     27\n",
+       "6    2020-6     29\n",
+       "7    2020-7     31\n",
+       "8    2020-8     31\n",
+       "9    2020-9     30\n",
+       "10   2020-10    31\n",
+       "11   2020-11    30\n",
+       "12   2020-12    30\n",
+       "13   2021-1     30\n",
+       "14   2021-2     24\n",
+       "Name: DIA, dtype: int64"
+      ]
+     },
+     "execution_count": 18,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_time.groupby(['MES','ANO+MES'])['DIA'].nunique()"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "e97f579e",
+   "id": "65efd8ed",
    "metadata": {},
    "outputs": [],
    "source": []