change to string nan for pandas replace

gnss_lib_py/parsers/nmea.py

@@ -151,7 +151,7 @@ def __init__(self, input_path, msg_types=None,
                                              .ffill()\
                                              .astype(float)
         pd_df["num_sats"] = pd_df["num_sats"].fillna(value=0).astype('int64')
-        pd_df = pd_df.fillna(value=np.nan).replace("",np.nan)
+        pd_df = pd_df.fillna(value=np.nan).replace("","nan")
         convert_dict={
                'num_sats' : np.int64,
                'gps_qual' : np.int64,