vtallen/testing123
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

first commit

Browse Source
This commit is contained in:
Vincent Allen
2026-05-21 08:40:24 -04:00
commit b084545275
711 changed files with 3659856 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
.ipynb_checkpoints/Untitled-checkpoint.ipynb Normal file
View File

@@ -0,0 +1,6 @@
{
"cells": [],
"metadata": {},
"nbformat": 4,
"nbformat_minor": 5
}

33760
.ipynb_checkpoints/clients_naics_data-checkpoint.csv Normal file
View File

File diff suppressed because it is too large Load Diff

BIN
.ipynb_checkpoints/comparison_graph-checkpoint.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 248 KiB

BIN
.ipynb_checkpoints/county_coverage_heat_map-checkpoint.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 479 KiB

BIN
.ipynb_checkpoints/county_heat_map-checkpoint.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 220 KiB

BIN
.ipynb_checkpoints/county_missing_heat_map-checkpoint.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 474 KiB

31
.ipynb_checkpoints/naics_data_csv-checkpoint.csv Normal file
View File

@@ -0,0 +1,31 @@
,EMP,EMP_F,EMP_IMP,EMP_IMP_F,ESTAB,ESTAB_F,FIRM,FIRM_F,GEO_ID,GEO_ID_F,NAME,NAICS2022,NAICS2022_F,NAICS2022_LABEL,PAYANN,PAYANN_F,PAYANN_IMP,PAYANN_IMP_F,PAYQTR1,PAYQTR1_F,RCPTOT,RCPTOT_F,RCPTOT_IMP,RCPTOT_IMP_F,TAXSTAT,TAXSTAT_LABEL,TYPOP,TYPOP_LABEL,YEAR,NAICS2022,ucgid
0,19792,,0,2,832,,614,,0400000US42,,Pennsylvania,21,,"Mining, quarrying, and oil and gas extraction",1784460,,0,2,445665,,44369022,,0,0,00,All establishments,00,All establishments,2022,21,0400000US42
1,30844,,0,1,917,,252,,0400000US42,,Pennsylvania,22,,Utilities,3538744,,0,0,1078106,,32259552,,0,0,00,All establishments,00,All establishments,2022,22,0400000US42
2,261765,,0,2,28792,,28433,,0400000US42,,Pennsylvania,23,,Construction,18896978,,0,0,4007954,,93996998,,0,1,00,All establishments,00,All establishments,2022,23,0400000US42
3,563035,,0,2,13377,,11984,,0400000US42,,Pennsylvania,31-33,,Manufacturing,36353647,,0,2,8835257,,262564512,,0,3,00,All establishments,00,All establishments,2022,31-33,0400000US42
4,252741,,0,3,12585,,9750,,0400000US42,,Pennsylvania,42,,Wholesale trade,21162185,,0,4,5607679,,411318239,,0,1,00,All establishments,00,All establishments,2022,42,0400000US42
5,202600,,0,3,10632,,8321,,0400000US42,,Pennsylvania,42,,Wholesale trade,15143956,,0,3,3728627,,259322722,,0,1,00,All establishments,10,"Merchant wholesalers, except manufacturers' sales branches and offices",2022,42,0400000US42
6,40420,,0,6,732,,330,,0400000US42,,Pennsylvania,42,,Wholesale trade,5294916,,0,7,1708821,,121491259,,0,1,00,All establishments,20,Manufacturers' sales branches and offices,2022,42,0400000US42
7,689445,,0,2,40165,,24552,,0400000US42,,Pennsylvania,44-45,,Retail trade,22798490,,0,2,5482917,,301886383,,0,2,00,All establishments,00,All establishments,2022,44-45,0400000US42
8,272785,,0,2,10858,,8798,,0400000US42,,Pennsylvania,48-49,107,Transportation and warehousing,13976679,,0,2,3337124,,42748848,,0,3,00,All establishments,00,All establishments,2022,48-49,0400000US42
9,107284,,0,3,5645,,2877,,0400000US42,,Pennsylvania,51,,Information,10240723,,0,3,2874268,,47462326,,0,3,00,All establishments,00,All establishments,2022,51,0400000US42
10,319889,,0,2,17692,,8906,,0400000US42,,Pennsylvania,52,115,Finance and insurance,32714999,,0,2,10239224,,270729542,,0,0,00,All establishments,00,All establishments,2022,52,0400000US42
11,77732,,0,2,11674,,8702,,0400000US42,,Pennsylvania,53,,Real estate and rental and leasing,4747740,,0,1,1182233,,27046613,,0,2,00,All establishments,00,All establishments,2022,53,0400000US42
12,391239,,0,2,30362,,27665,,0400000US42,,Pennsylvania,54,,"Professional, scientific, and technical services",35025855,,0,2,8635386,,92640000,,0,3,00,All establishments,00,All establishments,2022,54,0400000US42
13,385749,,0,2,30161,,27552,,0400000US42,,Pennsylvania,54,,"Professional, scientific, and technical services",34609148,,0,2,8540229,,91663124,,0,3,T,Establishments subject to federal income tax,00,All establishments,2022,54,0400000US42
14,5490,,0,2,201,,118,,0400000US42,,Pennsylvania,54,,"Professional, scientific, and technical services",416707,,0,1,95157,,976876,,0,3,Y,Establishments exempt from federal income tax,00,All establishments,2022,54,0400000US42
15,243649,,0,3,2356,,1369,,0400000US42,,Pennsylvania,55,,Management of companies and enterprises,22253581,,0,3,6614090,,2963566,,0,3,00,All establishments,00,All establishments,2022,55,0400000US42
16,362904,,0,2,16769,,14161,,0400000US42,,Pennsylvania,56,,Administrative and support and waste management and remediation services,16073022,,0,2,3744495,,41378855,,0,3,00,All establishments,00,All establishments,2022,56,0400000US42
17,26627,,0,1,2860,,2720,,0400000US42,,Pennsylvania,61,117,Educational services,957785,,0,1,222531,,2767178,,0,2,00,All establishments,00,All establishments,2022,61,0400000US42
18,18878,,0,2,2270,,2163,,0400000US42,,Pennsylvania,61,117,Educational services,600188,,0,2,137158,,1844599,,0,3,T,Establishments subject to federal income tax,00,All establishments,2022,61,0400000US42
19,7749,,0,1,590,,558,,0400000US42,,Pennsylvania,61,117,Educational services,357597,,0,1,85373,,922579,,0,1,Y,Establishments exempt from federal income tax,00,All establishments,2022,61,0400000US42
20,1109625,,0,2,39375,,23841,,0400000US42,,Pennsylvania,62,,Health care and social assistance,61823458,,0,2,14423975,,154216328,,0,4,00,All establishments,00,All establishments,2022,62,0400000US42
21,585444,,0,2,30646,,20539,,0400000US42,,Pennsylvania,62,,Health care and social assistance,32274263,,0,2,7491659,,74148150,,0,4,T,Establishments subject to federal income tax,00,All establishments,2022,62,0400000US42
22,524181,,0,3,8729,,3414,,0400000US42,,Pennsylvania,62,,Health care and social assistance,29549195,,0,3,6932316,,80068178,,0,4,Y,Establishments exempt from federal income tax,00,All establishments,2022,62,0400000US42
23,100712,,0,1,5076,,4693,,0400000US42,,Pennsylvania,71,,"Arts, entertainment, and recreation",4172509,,0,0,881024,,13079147,,0,1,00,All establishments,00,All establishments,2022,71,0400000US42
24,69561,,0,2,3762,,3501,,0400000US42,,Pennsylvania,71,,"Arts, entertainment, and recreation",3218925,,0,1,684522,,10360135,,0,1,T,Establishments subject to federal income tax,00,All establishments,2022,71,0400000US42
25,31151,,0,1,1314,,1194,,0400000US42,,Pennsylvania,71,,"Arts, entertainment, and recreation",953584,,0,0,196502,,2719012,,0,0,Y,Establishments exempt from federal income tax,00,All establishments,2022,71,0400000US42
26,462379,,0,1,29042,,23281,,0400000US42,,Pennsylvania,72,,Accommodation and food services,9724441,,0,1,2193189,,34888736,,0,2,00,All establishments,00,All establishments,2022,72,0400000US42
27,162373,,0,1,26781,,24498,,0400000US42,,Pennsylvania,81,119,Other services (except public administration),6204798,,0,1,1420492,,40428262,,0,0,00,All establishments,00,All establishments,2022,81,0400000US42
28,122551,,0,1,22075,,19955,,0400000US42,,Pennsylvania,81,119,Other services (except public administration),4422834,,0,1,1003809,,14823996,,0,2,T,Establishments subject to federal income tax,00,All establishments,2022,81,0400000US42
29,39822,,0,0,4706,,4548,,0400000US42,,Pennsylvania,81,119,Other services (except public administration),1781964,,0,0,416683,,25604266,,0,0,Y,Establishments exempt from federal income tax,00,All establishments,2022,81,0400000US42
1 EMP EMP_F EMP_IMP EMP_IMP_F ESTAB ESTAB_F FIRM FIRM_F GEO_ID GEO_ID_F NAME NAICS2022 NAICS2022_F NAICS2022_LABEL PAYANN PAYANN_F PAYANN_IMP PAYANN_IMP_F PAYQTR1 PAYQTR1_F RCPTOT RCPTOT_F RCPTOT_IMP RCPTOT_IMP_F TAXSTAT TAXSTAT_LABEL TYPOP TYPOP_LABEL YEAR NAICS2022 ucgid
2 0 19792 0 2 832 614 0400000US42 Pennsylvania 21 Mining, quarrying, and oil and gas extraction 1784460 0 2 445665 44369022 0 0 00 All establishments 00 All establishments 2022 21 0400000US42
3 1 30844 0 1 917 252 0400000US42 Pennsylvania 22 Utilities 3538744 0 0 1078106 32259552 0 0 00 All establishments 00 All establishments 2022 22 0400000US42
4 2 261765 0 2 28792 28433 0400000US42 Pennsylvania 23 Construction 18896978 0 0 4007954 93996998 0 1 00 All establishments 00 All establishments 2022 23 0400000US42
5 3 563035 0 2 13377 11984 0400000US42 Pennsylvania 31-33 Manufacturing 36353647 0 2 8835257 262564512 0 3 00 All establishments 00 All establishments 2022 31-33 0400000US42
6 4 252741 0 3 12585 9750 0400000US42 Pennsylvania 42 Wholesale trade 21162185 0 4 5607679 411318239 0 1 00 All establishments 00 All establishments 2022 42 0400000US42
7 5 202600 0 3 10632 8321 0400000US42 Pennsylvania 42 Wholesale trade 15143956 0 3 3728627 259322722 0 1 00 All establishments 10 Merchant wholesalers, except manufacturers' sales branches and offices 2022 42 0400000US42
8 6 40420 0 6 732 330 0400000US42 Pennsylvania 42 Wholesale trade 5294916 0 7 1708821 121491259 0 1 00 All establishments 20 Manufacturers' sales branches and offices 2022 42 0400000US42
9 7 689445 0 2 40165 24552 0400000US42 Pennsylvania 44-45 Retail trade 22798490 0 2 5482917 301886383 0 2 00 All establishments 00 All establishments 2022 44-45 0400000US42
10 8 272785 0 2 10858 8798 0400000US42 Pennsylvania 48-49 107 Transportation and warehousing 13976679 0 2 3337124 42748848 0 3 00 All establishments 00 All establishments 2022 48-49 0400000US42
11 9 107284 0 3 5645 2877 0400000US42 Pennsylvania 51 Information 10240723 0 3 2874268 47462326 0 3 00 All establishments 00 All establishments 2022 51 0400000US42
12 10 319889 0 2 17692 8906 0400000US42 Pennsylvania 52 115 Finance and insurance 32714999 0 2 10239224 270729542 0 0 00 All establishments 00 All establishments 2022 52 0400000US42
13 11 77732 0 2 11674 8702 0400000US42 Pennsylvania 53 Real estate and rental and leasing 4747740 0 1 1182233 27046613 0 2 00 All establishments 00 All establishments 2022 53 0400000US42
14 12 391239 0 2 30362 27665 0400000US42 Pennsylvania 54 Professional, scientific, and technical services 35025855 0 2 8635386 92640000 0 3 00 All establishments 00 All establishments 2022 54 0400000US42
15 13 385749 0 2 30161 27552 0400000US42 Pennsylvania 54 Professional, scientific, and technical services 34609148 0 2 8540229 91663124 0 3 T Establishments subject to federal income tax 00 All establishments 2022 54 0400000US42
16 14 5490 0 2 201 118 0400000US42 Pennsylvania 54 Professional, scientific, and technical services 416707 0 1 95157 976876 0 3 Y Establishments exempt from federal income tax 00 All establishments 2022 54 0400000US42
17 15 243649 0 3 2356 1369 0400000US42 Pennsylvania 55 Management of companies and enterprises 22253581 0 3 6614090 2963566 0 3 00 All establishments 00 All establishments 2022 55 0400000US42
18 16 362904 0 2 16769 14161 0400000US42 Pennsylvania 56 Administrative and support and waste management and remediation services 16073022 0 2 3744495 41378855 0 3 00 All establishments 00 All establishments 2022 56 0400000US42
19 17 26627 0 1 2860 2720 0400000US42 Pennsylvania 61 117 Educational services 957785 0 1 222531 2767178 0 2 00 All establishments 00 All establishments 2022 61 0400000US42
20 18 18878 0 2 2270 2163 0400000US42 Pennsylvania 61 117 Educational services 600188 0 2 137158 1844599 0 3 T Establishments subject to federal income tax 00 All establishments 2022 61 0400000US42
21 19 7749 0 1 590 558 0400000US42 Pennsylvania 61 117 Educational services 357597 0 1 85373 922579 0 1 Y Establishments exempt from federal income tax 00 All establishments 2022 61 0400000US42
22 20 1109625 0 2 39375 23841 0400000US42 Pennsylvania 62 Health care and social assistance 61823458 0 2 14423975 154216328 0 4 00 All establishments 00 All establishments 2022 62 0400000US42
23 21 585444 0 2 30646 20539 0400000US42 Pennsylvania 62 Health care and social assistance 32274263 0 2 7491659 74148150 0 4 T Establishments subject to federal income tax 00 All establishments 2022 62 0400000US42
24 22 524181 0 3 8729 3414 0400000US42 Pennsylvania 62 Health care and social assistance 29549195 0 3 6932316 80068178 0 4 Y Establishments exempt from federal income tax 00 All establishments 2022 62 0400000US42
25 23 100712 0 1 5076 4693 0400000US42 Pennsylvania 71 Arts, entertainment, and recreation 4172509 0 0 881024 13079147 0 1 00 All establishments 00 All establishments 2022 71 0400000US42
26 24 69561 0 2 3762 3501 0400000US42 Pennsylvania 71 Arts, entertainment, and recreation 3218925 0 1 684522 10360135 0 1 T Establishments subject to federal income tax 00 All establishments 2022 71 0400000US42
27 25 31151 0 1 1314 1194 0400000US42 Pennsylvania 71 Arts, entertainment, and recreation 953584 0 0 196502 2719012 0 0 Y Establishments exempt from federal income tax 00 All establishments 2022 71 0400000US42
28 26 462379 0 1 29042 23281 0400000US42 Pennsylvania 72 Accommodation and food services 9724441 0 1 2193189 34888736 0 2 00 All establishments 00 All establishments 2022 72 0400000US42
29 27 162373 0 1 26781 24498 0400000US42 Pennsylvania 81 119 Other services (except public administration) 6204798 0 1 1420492 40428262 0 0 00 All establishments 00 All establishments 2022 81 0400000US42
30 28 122551 0 1 22075 19955 0400000US42 Pennsylvania 81 119 Other services (except public administration) 4422834 0 1 1003809 14823996 0 2 T Establishments subject to federal income tax 00 All establishments 2022 81 0400000US42
31 29 39822 0 0 4706 4548 0400000US42 Pennsylvania 81 119 Other services (except public administration) 1781964 0 0 416683 25604266 0 0 Y Establishments exempt from federal income tax 00 All establishments 2022 81 0400000US42

33976
.ipynb_checkpoints/naics_tagged_client_list-checkpoint.csv Normal file
View File

File diff suppressed because it is too large Load Diff

1
.ipynb_checkpoints/pa_sbdc_chart_unified_axis-checkpoint.json Normal file
View File

File diff suppressed because one or more lines are too long

27543
.ipynb_checkpoints/section1_1-checkpoint.ipynb Normal file
View File

File diff suppressed because one or more lines are too long

624
.ipynb_checkpoints/section1_2-checkpoint.ipynb Normal file
View File

@@ -0,0 +1,624 @@
{
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"id": "036980a4-7309-4ea4-8dac-e8901c4525cf",
"metadata": {},
"outputs": [],
"source": [
"import pandas as pd"
]
},
{
"cell_type": "markdown",
"id": "ba89b925-5fc4-41bc-8b5e-81e53cfe1e4e",
"metadata": {},
"source": [
"# Getting the client counseling sessions data\n",
"---\n",
"The data can be obtained from this filter. You will need to break it up into smaller chunks and export them into the folder client_counseling\n",
"https://pasbdc.neoserra.com/activity/list/10?__formid=10&remove=&savename=&sort=DATE&sortdir=DESC&expr=&field_1=DATE&opt_auto_1=pfy&field_2=CLI_TYPE2&opt_2=&opt_2=AC&opt_2=IC&field_3=F_CENTER_ID&opt_3=2805&opt_3=2790&opt_3=2782&opt_3=2784&opt_3=2806&opt_3=2789&opt_3=4491&opt_3=2783&opt_3=2807&opt_3=2809&opt_3=2788&opt_3=2780&opt_3=2808&opt_3=2786&opt_3=2785&opt_3=2787&opt_3=2791&opt_3=2781&field_4=&sortdir=DESC\n",
"\n",
"Then combine the csvs into one big file"
]
},
{
"cell_type": "code",
"execution_count": 54,
"id": "7c3bf1e4-57bc-419b-909d-fca0043c1df7",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"<class 'pandas.core.frame.DataFrame'>\n",
"RangeIndex: 828614 entries, 0 to 828613\n",
"Data columns (total 9 columns):\n",
" # Column Non-Null Count Dtype \n",
"--- ------ -------------- ----- \n",
" 0 Session Date 828614 non-null object \n",
" 1 Client 828614 non-null object \n",
" 2 Client ID 828614 non-null object \n",
" 3 Counselor 823892 non-null object \n",
" 4 Session Type 828611 non-null object \n",
" 5 Contact Type 828614 non-null object \n",
" 6 Center 828614 non-null object \n",
" 7 Prep+Contact 828614 non-null float64\n",
" 8 Total Hours 828614 non-null float64\n",
"dtypes: float64(2), object(7)\n",
"memory usage: 56.9+ MB\n"
]
}
],
"source": [
"sessions_df = pd.read_csv('counselling_sessions_fy2225.csv')\n",
"sessions_df.info()"
]
},
{
"cell_type": "markdown",
"id": "acb44859-ef4f-4dcb-82a9-80e5c30ce778",
"metadata": {},
"source": [
"# Get the unique clients list\n",
"---\n",
"Download and load the unique clients list with these columns\n",
"https://pasbdc.neoserra.com/clients?__formid=3&remove=&savename=&sort=CLIENT_ID&sortdir=ASC&expr=&field_1=REVIEWID&opt_1=13213656&field_2=&sortdir=ASC"
]
},
{
"cell_type": "code",
"execution_count": 59,
"id": "4563ea40-4d94-41c5-86f4-6d620a45c1de",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>Unnamed: 0</th>\n",
" <th>Client ID</th>\n",
" <th>Client</th>\n",
" <th>Primary Contact</th>\n",
" <th>Last Counseling</th>\n",
" <th>Phone</th>\n",
" <th>Email</th>\n",
" <th>Physical Address</th>\n",
" <th>Physical Address County</th>\n",
" <th>Physical Address State</th>\n",
" <th>Primary NAICS</th>\n",
" <th>NAICs</th>\n",
" <th>NAICS_2</th>\n",
" <th>PA NAICs Code Percentage</th>\n",
" <th>PASBDC NAICs Code Percentage</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>0</td>\n",
" <td>WD04170</td>\n",
" <td>\\tProinnov@ LLC (WD04170)</td>\n",
" <td>Jardenson Castro</td>\n",
" <td>9/9/2025 12:00 AM</td>\n",
" <td>(267) 748-4465</td>\n",
" <td>JardensonC@ICLOUD.com</td>\n",
" <td>6752 Oakland St.</td>\n",
" <td>Philadelphia</td>\n",
" <td>Pennsylvania</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>0.0</td>\n",
" <td>0.000000</td>\n",
" <td>14.915377</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>1</td>\n",
" <td>WD02759</td>\n",
" <td>\"C.J.A.\"/ Crawley Jones and Allen real estate...</td>\n",
" <td>mark crawley</td>\n",
" <td>10/20/2025 12:00 AM</td>\n",
" <td>(215) 290-9828</td>\n",
" <td>mrkcrawley@gmail.com</td>\n",
" <td>673 Rively ave</td>\n",
" <td>Delaware</td>\n",
" <td>Pennsylvania</td>\n",
" <td>531390 - Other Activities Related to Real Esta...</td>\n",
" <td>531390-OtherActivitiesRelatedtoRealEstate\\r\\r\\...</td>\n",
" <td>53.0</td>\n",
" <td>2.510127</td>\n",
" <td>2.688026</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>2</td>\n",
" <td>PS018402</td>\n",
" <td>Anjie's Cleaning Bees (PS018402)</td>\n",
" <td>Anjelica Gonzez</td>\n",
" <td>10/14/2024 12:00 AM</td>\n",
" <td>(717) 521-3625</td>\n",
" <td>anjelicagonzalez2001@gmail.com</td>\n",
" <td>1129 High St</td>\n",
" <td>Lycoming</td>\n",
" <td>Pennsylvania</td>\n",
" <td>561720 - Janitorial Services \\r\\r\\n</td>\n",
" <td>561720-JanitorialServices\\r\\r\\n\\r\\r\\n</td>\n",
" <td>56.0</td>\n",
" <td>3.605647</td>\n",
" <td>4.344285</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>3</td>\n",
" <td>C8538</td>\n",
" <td>BRENIMAN PROPERTIES, LLC (C8538)</td>\n",
" <td>RYAN BRENIMAN</td>\n",
" <td>10/17/2025 12:00 AM</td>\n",
" <td>NaN</td>\n",
" <td>r_breniman@yahoo.com</td>\n",
" <td>147 Heeter Rd</td>\n",
" <td>Clarion</td>\n",
" <td>Pennsylvania</td>\n",
" <td>531120 - Lessors of Nonresidential Buildings (...</td>\n",
" <td>531120-LessorsofNonresidentialBuildings(except...</td>\n",
" <td>53.0</td>\n",
" <td>2.510127</td>\n",
" <td>2.688026</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>4</td>\n",
" <td>BU016079</td>\n",
" <td>Civil War Cider Co., Inc. (BU016079)</td>\n",
" <td>Robert Antanitis, II</td>\n",
" <td>10/21/2024 12:00 AM</td>\n",
" <td>(570) 523-3414</td>\n",
" <td>rob@civilwarcider.com</td>\n",
" <td>606 Market St.</td>\n",
" <td>Union</td>\n",
" <td>Pennsylvania</td>\n",
" <td>312130 - Wineries \\r\\r\\n</td>\n",
" <td>312130-Wineries\\r\\r\\n\\r\\r\\n</td>\n",
" <td>31.0</td>\n",
" <td>2.876304</td>\n",
" <td>4.923522</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" Unnamed: 0 Client ID Client \\\n",
"0 0 WD04170 \\tProinnov@ LLC (WD04170) \n",
"1 1 WD02759 \"C.J.A.\"/ Crawley Jones and Allen real estate... \n",
"2 2 PS018402 Anjie's Cleaning Bees (PS018402) \n",
"3 3 C8538 BRENIMAN PROPERTIES, LLC (C8538) \n",
"4 4 BU016079 Civil War Cider Co., Inc. (BU016079) \n",
"\n",
" Primary Contact Last Counseling Phone \\\n",
"0 Jardenson Castro 9/9/2025 12:00 AM (267) 748-4465 \n",
"1 mark crawley 10/20/2025 12:00 AM (215) 290-9828 \n",
"2 Anjelica Gonzez 10/14/2024 12:00 AM (717) 521-3625 \n",
"3 RYAN BRENIMAN 10/17/2025 12:00 AM NaN \n",
"4 Robert Antanitis, II 10/21/2024 12:00 AM (570) 523-3414 \n",
"\n",
" Email Physical Address Physical Address County \\\n",
"0 JardensonC@ICLOUD.com 6752 Oakland St. Philadelphia \n",
"1 mrkcrawley@gmail.com 673 Rively ave Delaware \n",
"2 anjelicagonzalez2001@gmail.com 1129 High St Lycoming \n",
"3 r_breniman@yahoo.com 147 Heeter Rd Clarion \n",
"4 rob@civilwarcider.com 606 Market St. Union \n",
"\n",
" Physical Address State Primary NAICS \\\n",
"0 Pennsylvania NaN \n",
"1 Pennsylvania 531390 - Other Activities Related to Real Esta... \n",
"2 Pennsylvania 561720 - Janitorial Services \\r\\r\\n \n",
"3 Pennsylvania 531120 - Lessors of Nonresidential Buildings (... \n",
"4 Pennsylvania 312130 - Wineries \\r\\r\\n \n",
"\n",
" NAICs NAICS_2 \\\n",
"0 NaN 0.0 \n",
"1 531390-OtherActivitiesRelatedtoRealEstate\\r\\r\\... 53.0 \n",
"2 561720-JanitorialServices\\r\\r\\n\\r\\r\\n 56.0 \n",
"3 531120-LessorsofNonresidentialBuildings(except... 53.0 \n",
"4 312130-Wineries\\r\\r\\n\\r\\r\\n 31.0 \n",
"\n",
" PA NAICs Code Percentage PASBDC NAICs Code Percentage \n",
"0 0.000000 14.915377 \n",
"1 2.510127 2.688026 \n",
"2 3.605647 4.344285 \n",
"3 2.510127 2.688026 \n",
"4 2.876304 4.923522 "
]
},
"execution_count": 59,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"client_df = pd.read_csv('naics_client_list_tagged.csv')\n",
"client_df.head()"
]
},
{
"cell_type": "code",
"execution_count": 60,
"id": "9307bc7a-8ba1-4c7f-b7eb-e2b5f6d4c8d7",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>Session Type</th>\n",
" <th>count</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>Administrative</td>\n",
" <td>476099</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>Follow-up</td>\n",
" <td>316915</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>Initial/New</td>\n",
" <td>17024</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>Close-out</td>\n",
" <td>14902</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>Impact</td>\n",
" <td>3671</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" Session Type count\n",
"0 Administrative 476099\n",
"1 Follow-up 316915\n",
"2 Initial/New 17024\n",
"3 Close-out 14902\n",
"4 Impact 3671"
]
},
"execution_count": 60,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"session_type_value_counts = sessions_df['Session Type'].value_counts().reset_index()\n",
"session_type_value_counts"
]
},
{
"cell_type": "code",
"execution_count": 83,
"id": "97465a49-4e13-4e23-88ba-6bd7d72c6520",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th>Session Type</th>\n",
" <th>Client ID</th>\n",
" <th>Center</th>\n",
" <th>Administrative</th>\n",
" <th>Close-out</th>\n",
" <th>Follow-up</th>\n",
" <th>Impact</th>\n",
" <th>Initial/New</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>D 14632</td>\n",
" <td>Duquesne University SBDC</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>2</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>S015040</td>\n",
" <td>The University of Scranton SBDC</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>7</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>00000043</td>\n",
" <td>TE - TEMPLE SBDC</td>\n",
" <td>1</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>00000052</td>\n",
" <td>WD - WIDENER SBDC</td>\n",
" <td>3</td>\n",
" <td>0</td>\n",
" <td>8</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>000002</td>\n",
" <td>Kutztown University SBDC</td>\n",
" <td>2</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
"Session Type Client ID Center Administrative \\\n",
"0 D 14632 Duquesne University SBDC 0 \n",
"1 S015040 The University of Scranton SBDC 0 \n",
"2 00000043 TE - TEMPLE SBDC 1 \n",
"3 00000052 WD - WIDENER SBDC 3 \n",
"4 000002 Kutztown University SBDC 2 \n",
"\n",
"Session Type Close-out Follow-up Impact Initial/New \n",
"0 0 2 0 1 \n",
"1 1 7 0 0 \n",
"2 0 0 0 0 \n",
"3 0 8 0 0 \n",
"4 0 0 0 0 "
]
},
"execution_count": 83,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"session_counts = sessions_df.groupby('Client ID')['Session Type'].value_counts()\n",
"client_centers = sessions_df.groupby('Client ID')['Center'].first()\n",
"\n",
"\n",
"unified_counts_df = session_counts.unstack(fill_value=0)\n",
"unified_counts_df['Center'] = client_centers\n",
"\n",
"\n",
"\n",
"all_cols = list(unified_counts_df.columns)\n",
"all_cols.remove('Center')\n",
"all_cols.insert(0, 'Center')\n",
"\n",
"unified_counts_df = unified_counts_df[all_cols].reset_index()\n",
"unified_counts_df.head()"
]
},
{
"cell_type": "code",
"execution_count": 78,
"id": "0eb344e3-a1e2-4ae8-b2ad-a40cc775eec9",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Center\n",
"University of Pittsburgh SBDC 11263\n",
"TE - TEMPLE SBDC 8851\n",
"Kutztown University SBDC 8454\n",
"WD - WIDENER SBDC 4452\n",
"The University of Scranton SBDC 3798\n",
"K - Kutztown SBDC 2608\n",
"PennWest University Clarion SBDC 2574\n",
"WI - WILKES SBDC 2387\n",
"LE - LEHIGH UNIVERSITY SBDC 2211\n",
"G - GANNON SBDC 1794\n",
"Penn State SBDC 1782\n",
"SH - SHIPPENSBURG SBDC 1751\n",
"Duquesne University SBDC 1604\n",
"Bucknell SBDC 1149\n",
"SF - ST. FRANCIS UNIVERSITY SBDC 1033\n",
"SV - ST. VINCENT COLLEGE SBDC 738\n",
"G - Meadville 186\n",
"SV - Fayette Outreach 176\n",
"G - Mercer 131\n",
"Indiana County 125\n",
"Clarion CARES Act 20\n",
"G - Warren 1\n",
"SC - Monroe Outreach 1\n",
"LE - Bucks County/Lehigh SBDC 1\n",
"Name: count, dtype: int64"
]
},
"execution_count": 78,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"unified_counts_df['Center'].value_counts()"
]
},
{
"cell_type": "code",
"execution_count": 86,
"id": "e38f57b8-f362-4bba-ae20-b858f2be4504",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Center\n",
"Pittsburgh 11263\n",
"Temple 8851\n",
"Kutztown University SBDC 8454\n",
"Widner 4452\n",
"Scranton 3798\n",
"Kutztown 2608\n",
"Clarion 2574\n",
"Wilkes 2387\n",
"Lehigh 2212\n",
"Gannon 1794\n",
"Penn State 1782\n",
"Shippensburg 1751\n",
"Duquesne 1604\n",
"Bucknell 1149\n",
"St. Francis 1033\n",
"St. Vincent 738\n",
"G - Meadville 186\n",
"SV - Fayette Outreach 176\n",
"G - Mercer 131\n",
"Indiana County 125\n",
"Clarion CARES Act 20\n",
"G - Warren 1\n",
"SC - Monroe Outreach 1\n",
"Name: count, dtype: int64"
]
},
"execution_count": 86,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"center_mapping = {\n",
" \"University of Pittsburgh SBDC\":\"Pittsburgh\",\n",
" \"TE - TEMPLE SBDC\":\"Temple\",\n",
" \"Kutztown University SBDC\": \"Kutztown\",\n",
" \"K - Kutztown SBDC\":\"Kutztown\",\n",
" \"WD - WIDENER SBDC\": \"Widner\",\n",
" \"The University of Scranton SBDC\": \"Scranton\",\n",
" \"PennWest University Clarion SBDC\":\"Clarion\",\n",
" \"WI - WILKES SBDC\":\"Wilkes\",\n",
" \"LE - LEHIGH UNIVERSITY SBDC\":\"Lehigh\",\n",
" \"G - GANNON SBDC\":\"Gannon\",\n",
" \"Penn State SBDC\":\"Penn State\",\n",
" \"SH - SHIPPENSBURG SBDC\":\"Shippensburg\",\n",
" \"Duquesne University SBDC\":\"Duquesne\",\n",
" \"Bucknell SBDC\":\"Bucknell\",\n",
" \"SF - ST. FRANCIS UNIVERSITY SBDC\": \"St. Francis\",\n",
" \"SV - ST. VINCENT COLLEGE SBDC\":\"St. Vincent\",\n",
" \"LE - Bucks County/Lehigh SBDC\":\"Lehigh\",\n",
"}\n",
"\n",
"unified_counts_df['Center'] = unified_counts_df['Center'].replace(center_mapping)\n",
"unified_counts_df['Center'].value_counts()"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "39772f4d-0c9b-40cb-ab7a-0d3e152bd7a3",
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"id": "369dddf9-4b1a-47ea-aa08-cc0404c0f60a",
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.13.7"
}
},
"nbformat": 4,
"nbformat_minor": 5
}

BIN
.ipynb_checkpoints/sector_table-checkpoint.jpg Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 75 KiB